modeling/20-PyTorch2ONNX.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import io\n",
    "import numpy as np\n",
    "\n",
    "from torch import nn\n",
    "import torch.utils.model_zoo as model_zoo\n",
    "import torch.onnx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.init as init"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils.google_utils import *\n",
    "from utils.layers import *\n",
    "from utils.parse_config import *\n",
    "\n",
    "ONNX_EXPORT = False\n",
    "\n",
    "\n",
    "def create_modules(module_defs, img_size, cfg):\n",
    "    # Constructs module list of layer blocks from module configuration in module_defs\n",
    "\n",
    "    img_size = [img_size] * 2 if isinstance(img_size, int) else img_size  # expand if necessary\n",
    "    _ = module_defs.pop(0)  # cfg training hyperparams (unused)\n",
    "    output_filters = [3]  # input channels\n",
    "    module_list = nn.ModuleList()\n",
    "    routs = []  # list of layers which rout to deeper layers\n",
    "    yolo_index = -1\n",
    "\n",
    "    for i, mdef in enumerate(module_defs):\n",
    "        modules = nn.Sequential()\n",
    "\n",
    "        if mdef['type'] == 'convolutional':\n",
    "            bn = mdef['batch_normalize']\n",
    "            filters = mdef['filters']\n",
    "            k = mdef['size']  # kernel size\n",
    "            stride = mdef['stride'] if 'stride' in mdef else (mdef['stride_y'], mdef['stride_x'])\n",
    "            if isinstance(k, int):  # single-size conv\n",
    "                modules.add_module('Conv2d', nn.Conv2d(in_channels=output_filters[-1],\n",
    "                                                       out_channels=filters,\n",
    "                                                       kernel_size=k,\n",
    "                                                       stride=stride,\n",
    "                                                       padding=k // 2 if mdef['pad'] else 0,\n",
    "                                                       groups=mdef['groups'] if 'groups' in mdef else 1,\n",
    "                                                       bias=not bn))\n",
    "            else:  # multiple-size conv\n",
    "                modules.add_module('MixConv2d', MixConv2d(in_ch=output_filters[-1],\n",
    "                                                          out_ch=filters,\n",
    "                                                          k=k,\n",
    "                                                          stride=stride,\n",
    "                                                          bias=not bn))\n",
    "\n",
    "            if bn:\n",
    "                modules.add_module('BatchNorm2d', nn.BatchNorm2d(filters, momentum=0.03, eps=1E-4))\n",
    "            else:\n",
    "                routs.append(i)  # detection output (goes into yolo layer)\n",
    "\n",
    "            if mdef['activation'] == 'leaky':  # activation study https://github.com/ultralytics/yolov3/issues/441\n",
    "                modules.add_module('activation', nn.LeakyReLU(0.1, inplace=True))\n",
    "            elif mdef['activation'] == 'swish':\n",
    "                modules.add_module('activation', Swish())\n",
    "            elif mdef['activation'] == 'mish':\n",
    "                modules.add_module('activation', Mish())\n",
    "            elif mdef['activation'] == 'relu':\n",
    "                modules.add_module('activation', nn.ReLU())\n",
    "\n",
    "        elif mdef['type'] == 'BatchNorm2d':\n",
    "            filters = output_filters[-1]\n",
    "            modules = nn.BatchNorm2d(filters, momentum=0.03, eps=1E-4)\n",
    "            if i == 0 and filters == 3:  # normalize RGB image\n",
    "                # imagenet mean and var https://pytorch.org/docs/stable/torchvision/models.html#classification\n",
    "                modules.running_mean = torch.tensor([0.485, 0.456, 0.406])\n",
    "                modules.running_var = torch.tensor([0.0524, 0.0502, 0.0506])\n",
    "\n",
    "        elif mdef['type'] == 'maxpool':\n",
    "            k = mdef['size']  # kernel size\n",
    "            stride = mdef['stride']\n",
    "            maxpool = nn.MaxPool2d(kernel_size=k, stride=stride, padding=(k - 1) // 2)\n",
    "            if k == 2 and stride == 1:  # yolov3-tiny\n",
    "                modules.add_module('ZeroPad2d', nn.ZeroPad2d((0, 1, 0, 1)))\n",
    "                modules.add_module('MaxPool2d', maxpool)\n",
    "            else:\n",
    "                modules = maxpool\n",
    "\n",
    "        elif mdef['type'] == 'upsample':\n",
    "            if ONNX_EXPORT:  # explicitly state size, avoid scale_factor\n",
    "                g = (yolo_index + 1) * 2 / 32  # gain\n",
    "                modules = nn.Upsample(size=tuple(int(x * g) for x in img_size))  # img_size = (320, 192)\n",
    "            else:\n",
    "                modules = nn.Upsample(scale_factor=mdef['stride'])\n",
    "\n",
    "        elif mdef['type'] == 'route':  # nn.Sequential() placeholder for 'route' layer\n",
    "            layers = mdef['layers']\n",
    "            filters = sum([output_filters[l + 1 if l > 0 else l] for l in layers])\n",
    "            routs.extend([i + l if l < 0 else l for l in layers])\n",
    "            modules = FeatureConcat(layers=layers)\n",
    "\n",
    "        elif mdef['type'] == 'shortcut':  # nn.Sequential() placeholder for 'shortcut' layer\n",
    "            layers = mdef['from']\n",
    "            filters = output_filters[-1]\n",
    "            routs.extend([i + l if l < 0 else l for l in layers])\n",
    "            modules = WeightedFeatureFusion(layers=layers, weight='weights_type' in mdef)\n",
    "\n",
    "        elif mdef['type'] == 'reorg3d':  # yolov3-spp-pan-scale\n",
    "            pass\n",
    "\n",
    "        elif mdef['type'] == 'yolo':\n",
    "            yolo_index += 1\n",
    "            stride = [32, 16, 8]  # P5, P4, P3 strides\n",
    "            if any(x in cfg for x in ['panet', 'yolov4', 'cd53']):  # stride order reversed\n",
    "                stride = list(reversed(stride))\n",
    "            layers = mdef['from'] if 'from' in mdef else []\n",
    "            modules = YOLOLayer(anchors=mdef['anchors'][mdef['mask']],  # anchor list\n",
    "                                nc=mdef['classes'],  # number of classes\n",
    "                                img_size=img_size,  # (416, 416)\n",
    "                                yolo_index=yolo_index,  # 0, 1, 2...\n",
    "                                layers=layers,  # output layers\n",
    "                                stride=stride[yolo_index])\n",
    "\n",
    "            # Initialize preceding Conv2d() bias (https://arxiv.org/pdf/1708.02002.pdf section 3.3)\n",
    "            try:\n",
    "                j = layers[yolo_index] if 'from' in mdef else -1\n",
    "                # If previous layer is a dropout layer, get the one before\n",
    "                if module_list[j].__class__.__name__ == 'Dropout':\n",
    "                    j -= 1\n",
    "                bias_ = module_list[j][0].bias  # shape(255,)\n",
    "                bias = bias_[:modules.no * modules.na].view(modules.na, -1)  # shape(3,85)\n",
    "                bias[:, 4] += -4.5  # obj\n",
    "                bias[:, 5:] += math.log(0.6 / (modules.nc - 0.99))  # cls (sigmoid(p) = 1/nc)\n",
    "                module_list[j][0].bias = torch.nn.Parameter(bias_, requires_grad=bias_.requires_grad)\n",
    "            except:\n",
    "                print('WARNING: smart bias initialization failure.')\n",
    "\n",
    "        elif mdef['type'] == 'dropout':\n",
    "            perc = float(mdef['probability'])\n",
    "            modules = nn.Dropout(p=perc)\n",
    "        else:\n",
    "            print('Warning: Unrecognized Layer Type: ' + mdef['type'])\n",
    "\n",
    "        # Register module list and number of output filters\n",
    "        module_list.append(modules)\n",
    "        output_filters.append(filters)\n",
    "\n",
    "    routs_binary = [False] * (i + 1)\n",
    "    for i in routs:\n",
    "        routs_binary[i] = True\n",
    "    return module_list, routs_binary\n",
    "\n",
    "\n",
    "class YOLOLayer(nn.Module):\n",
    "    def __init__(self, anchors, nc, img_size, yolo_index, layers, stride):\n",
    "        super(YOLOLayer, self).__init__()\n",
    "        self.anchors = torch.Tensor(anchors)\n",
    "        self.index = yolo_index  # index of this layer in layers\n",
    "        self.layers = layers  # model output layer indices\n",
    "        self.stride = stride  # layer stride\n",
    "        self.nl = len(layers)  # number of output layers (3)\n",
    "        self.na = len(anchors)  # number of anchors (3)\n",
    "        self.nc = nc  # number of classes (80)\n",
    "        self.no = nc + 5  # number of outputs (85)\n",
    "        self.nx, self.ny, self.ng = 0, 0, 0  # initialize number of x, y gridpoints\n",
    "        self.anchor_vec = self.anchors / self.stride\n",
    "        self.anchor_wh = self.anchor_vec.view(1, self.na, 1, 1, 2)\n",
    "\n",
    "        if ONNX_EXPORT:\n",
    "            self.training = False\n",
    "            self.create_grids((img_size[1] // stride, img_size[0] // stride))  # number x, y grid points\n",
    "\n",
    "    def create_grids(self, ng=(13, 13), device='cpu'):\n",
    "        self.nx, self.ny = ng  # x and y grid size\n",
    "        self.ng = torch.tensor(ng, dtype=torch.float)\n",
    "\n",
    "        # build xy offsets\n",
    "        if not self.training:\n",
    "            yv, xv = torch.meshgrid([torch.arange(self.ny, device=device), torch.arange(self.nx, device=device)])\n",
    "            self.grid = torch.stack((xv, yv), 2).view((1, 1, self.ny, self.nx, 2)).float()\n",
    "\n",
    "        if self.anchor_vec.device != device:\n",
    "            self.anchor_vec = self.anchor_vec.to(device)\n",
    "            self.anchor_wh = self.anchor_wh.to(device)\n",
    "\n",
    "    def forward(self, p, out):\n",
    "        ASFF = False  # https://arxiv.org/abs/1911.09516\n",
    "        if ASFF:\n",
    "            i, n = self.index, self.nl  # index in layers, number of layers\n",
    "            p = out[self.layers[i]]\n",
    "            bs, _, ny, nx = p.shape  # bs, 255, 13, 13\n",
    "            if (self.nx, self.ny) != (nx, ny):\n",
    "                self.create_grids((nx, ny), p.device)\n",
    "\n",
    "            # outputs and weights\n",
    "            # w = F.softmax(p[:, -n:], 1)  # normalized weights\n",
    "            w = torch.sigmoid(p[:, -n:]) * (2 / n)  # sigmoid weights (faster)\n",
    "            # w = w / w.sum(1).unsqueeze(1)  # normalize across layer dimension\n",
    "\n",
    "            # weighted ASFF sum\n",
    "            p = out[self.layers[i]][:, :-n] * w[:, i:i + 1]\n",
    "            for j in range(n):\n",
    "                if j != i:\n",
    "                    p += w[:, j:j + 1] * \\\n",
    "                         F.interpolate(out[self.layers[j]][:, :-n], size=[ny, nx], mode='bilinear', align_corners=False)\n",
    "\n",
    "        elif ONNX_EXPORT:\n",
    "            bs = 1  # batch size\n",
    "        else:\n",
    "            bs, _, ny, nx = p.shape  # bs, 255, 13, 13\n",
    "            if (self.nx, self.ny) != (nx, ny):\n",
    "                self.create_grids((nx, ny), p.device)\n",
    "\n",
    "        # p.view(bs, 255, 13, 13) -- > (bs, 3, 13, 13, 85)  # (bs, anchors, grid, grid, classes + xywh)\n",
    "        p = p.view(bs, self.na, self.no, self.ny, self.nx).permute(0, 1, 3, 4, 2).contiguous()  # prediction\n",
    "\n",
    "        if self.training:\n",
    "            return p\n",
    "\n",
    "        elif ONNX_EXPORT:\n",
    "            # Avoid broadcasting for ANE operations\n",
    "            m = self.na * self.nx * self.ny\n",
    "            ng = 1. / self.ng.repeat(m, 1)\n",
    "            grid = self.grid.repeat(1, self.na, 1, 1, 1).view(m, 2)\n",
    "            anchor_wh = self.anchor_wh.repeat(1, 1, self.nx, self.ny, 1).view(m, 2) * ng\n",
    "\n",
    "            p = p.view(m, self.no)\n",
    "            xy = torch.sigmoid(p[:, 0:2]) + grid  # x, y\n",
    "            wh = torch.exp(p[:, 2:4]) * anchor_wh  # width, height\n",
    "            p_cls = torch.sigmoid(p[:, 4:5]) if self.nc == 1 else \\\n",
    "                torch.sigmoid(p[:, 5:self.no]) * torch.sigmoid(p[:, 4:5])  # conf\n",
    "            return p_cls, xy * ng, wh\n",
    "\n",
    "        else:  # inference\n",
    "            io = p.clone()  # inference output\n",
    "            io[..., :2] = torch.sigmoid(io[..., :2]) + self.grid  # xy\n",
    "            io[..., 2:4] = torch.exp(io[..., 2:4]) * self.anchor_wh  # wh yolo method\n",
    "            io[..., :4] *= self.stride\n",
    "            torch.sigmoid_(io[..., 4:])\n",
    "            return io.view(bs, -1, self.no), p  # view [1, 3, 13, 13, 85] as [1, 507, 85]\n",
    "\n",
    "\n",
    "class Darknet(nn.Module):\n",
    "    # YOLOv3 object detection model\n",
    "\n",
    "    def __init__(self, cfg, img_size=(416, 416), verbose=False):\n",
    "        super(Darknet, self).__init__()\n",
    "\n",
    "        self.module_defs = parse_model_cfg(cfg)\n",
    "        self.module_list, self.routs = create_modules(self.module_defs, img_size, cfg)\n",
    "        self.yolo_layers = get_yolo_layers(self)\n",
    "        # torch_utils.initialize_weights(self)\n",
    "\n",
    "        # Darknet Header https://github.com/AlexeyAB/darknet/issues/2914#issuecomment-496675346\n",
    "        self.version = np.array([0, 2, 5], dtype=np.int32)  # (int32) version info: major, minor, revision\n",
    "        self.seen = np.array([0], dtype=np.int64)  # (int64) number of images seen during training\n",
    "        self.info(verbose) if not ONNX_EXPORT else None  # print model description\n",
    "\n",
    "    def forward(self, x, augment=False, verbose=False):\n",
    "\n",
    "        if not augment:\n",
    "            return self.forward_once(x)\n",
    "        else:  # Augment images (inference and test only) https://github.com/ultralytics/yolov3/issues/931\n",
    "            img_size = x.shape[-2:]  # height, width\n",
    "            s = [0.83, 0.67]  # scales\n",
    "            y = []\n",
    "            for i, xi in enumerate((x,\n",
    "                                    torch_utils.scale_img(x.flip(3), s[0], same_shape=False),  # flip-lr and scale\n",
    "                                    torch_utils.scale_img(x, s[1], same_shape=False),  # scale\n",
    "                                    )):\n",
    "                # cv2.imwrite('img%g.jpg' % i, 255 * xi[0].numpy().transpose((1, 2, 0))[:, :, ::-1])\n",
    "                y.append(self.forward_once(xi)[0])\n",
    "\n",
    "            y[1][..., :4] /= s[0]  # scale\n",
    "            y[1][..., 0] = img_size[1] - y[1][..., 0]  # flip lr\n",
    "            y[2][..., :4] /= s[1]  # scale\n",
    "\n",
    "            # for i, yi in enumerate(y):  # coco small, medium, large = < 32**2 < 96**2 <\n",
    "            #     area = yi[..., 2:4].prod(2)[:, :, None]\n",
    "            #     if i == 1:\n",
    "            #         yi *= (area < 96. ** 2).float()\n",
    "            #     elif i == 2:\n",
    "            #         yi *= (area > 32. ** 2).float()\n",
    "            #     y[i] = yi\n",
    "\n",
    "            y = torch.cat(y, 1)\n",
    "            return y, None\n",
    "\n",
    "    def forward_once(self, x, augment=False, verbose=False):\n",
    "        img_size = x.shape[-2:]  # height, width\n",
    "        yolo_out, out = [], []\n",
    "        if verbose:\n",
    "            print('0', x.shape)\n",
    "            str = ''\n",
    "\n",
    "        # Augment images (inference and test only)\n",
    "        if augment:  # https://github.com/ultralytics/yolov3/issues/931\n",
    "            nb = x.shape[0]  # batch size\n",
    "            s = [0.83, 0.67]  # scales\n",
    "            x = torch.cat((x,\n",
    "                           torch_utils.scale_img(x.flip(3), s[0]),  # flip-lr and scale\n",
    "                           torch_utils.scale_img(x, s[1]),  # scale\n",
    "                           ), 0)\n",
    "\n",
    "        for i, module in enumerate(self.module_list):\n",
    "            name = module.__class__.__name__\n",
    "            if name in ['WeightedFeatureFusion', 'FeatureConcat']:  # sum, concat\n",
    "                if verbose:\n",
    "                    l = [i - 1] + module.layers  # layers\n",
    "                    sh = [list(x.shape)] + [list(out[i].shape) for i in module.layers]  # shapes\n",
    "                    str = ' >> ' + ' + '.join(['layer %g %s' % x for x in zip(l, sh)])\n",
    "                x = module(x, out)  # WeightedFeatureFusion(), FeatureConcat()\n",
    "            elif name == 'YOLOLayer':\n",
    "                yolo_out.append(module(x, out))\n",
    "            else:  # run module directly, i.e. mtype = 'convolutional', 'upsample', 'maxpool', 'batchnorm2d' etc.\n",
    "                x = module(x)\n",
    "\n",
    "            out.append(x if self.routs[i] else [])\n",
    "            if verbose:\n",
    "                print('%g/%g %s -' % (i, len(self.module_list), name), list(x.shape), str)\n",
    "                str = ''\n",
    "\n",
    "        if self.training:  # train\n",
    "            return yolo_out\n",
    "        elif ONNX_EXPORT:  # export\n",
    "            x = [torch.cat(x, 0) for x in zip(*yolo_out)]\n",
    "            return x[0], torch.cat(x[1:3], 1)  # scores, boxes: 3780x80, 3780x4\n",
    "        else:  # inference or test\n",
    "            x, p = zip(*yolo_out)  # inference output, training output\n",
    "            x = torch.cat(x, 1)  # cat yolo outputs\n",
    "            if augment:  # de-augment results\n",
    "                x = torch.split(x, nb, dim=0)\n",
    "                x[1][..., :4] /= s[0]  # scale\n",
    "                x[1][..., 0] = img_size[1] - x[1][..., 0]  # flip lr\n",
    "                x[2][..., :4] /= s[1]  # scale\n",
    "                x = torch.cat(x, 1)\n",
    "            return x, p\n",
    "\n",
    "    def fuse(self):\n",
    "        # Fuse Conv2d + BatchNorm2d layers throughout model\n",
    "        print('Fusing layers...')\n",
    "        fused_list = nn.ModuleList()\n",
    "        for a in list(self.children())[0]:\n",
    "            if isinstance(a, nn.Sequential):\n",
    "                for i, b in enumerate(a):\n",
    "                    if isinstance(b, nn.modules.batchnorm.BatchNorm2d):\n",
    "                        # fuse this bn layer with the previous conv2d layer\n",
    "                        conv = a[i - 1]\n",
    "                        fused = torch_utils.fuse_conv_and_bn(conv, b)\n",
    "                        a = nn.Sequential(fused, *list(a.children())[i + 1:])\n",
    "                        break\n",
    "            fused_list.append(a)\n",
    "        self.module_list = fused_list\n",
    "        self.info() if not ONNX_EXPORT else None  # yolov3-spp reduced from 225 to 152 layers\n",
    "\n",
    "    def info(self, verbose=False):\n",
    "        torch_utils.model_info(self, verbose)\n",
    "\n",
    "\n",
    "def get_yolo_layers(model):\n",
    "    return [i for i, m in enumerate(model.module_list) if m.__class__.__name__ == 'YOLOLayer']  # [89, 101, 113]\n",
    "\n",
    "\n",
    "def load_darknet_weights(self, weights, cutoff=-1):\n",
    "    # Parses and loads the weights stored in 'weights'\n",
    "\n",
    "    # Establish cutoffs (load layers between 0 and cutoff. if cutoff = -1 all are loaded)\n",
    "    file = Path(weights).name\n",
    "    if file == 'darknet53.conv.74':\n",
    "        cutoff = 75\n",
    "    elif file == 'yolov3-tiny.conv.15':\n",
    "        cutoff = 15\n",
    "\n",
    "    # Read weights file\n",
    "    with open(weights, 'rb') as f:\n",
    "        # Read Header https://github.com/AlexeyAB/darknet/issues/2914#issuecomment-496675346\n",
    "        self.version = np.fromfile(f, dtype=np.int32, count=3)  # (int32) version info: major, minor, revision\n",
    "        self.seen = np.fromfile(f, dtype=np.int64, count=1)  # (int64) number of images seen during training\n",
    "\n",
    "        weights = np.fromfile(f, dtype=np.float32)  # the rest are weights\n",
    "\n",
    "    ptr = 0\n",
    "    for i, (mdef, module) in enumerate(zip(self.module_defs[:cutoff], self.module_list[:cutoff])):\n",
    "        if mdef['type'] == 'convolutional':\n",
    "            conv = module[0]\n",
    "            if mdef['batch_normalize']:\n",
    "                # Load BN bias, weights, running mean and running variance\n",
    "                bn = module[1]\n",
    "                nb = bn.bias.numel()  # number of biases\n",
    "                # Bias\n",
    "                bn.bias.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.bias))\n",
    "                ptr += nb\n",
    "                # Weight\n",
    "                bn.weight.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.weight))\n",
    "                ptr += nb\n",
    "                # Running Mean\n",
    "                bn.running_mean.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.running_mean))\n",
    "                ptr += nb\n",
    "                # Running Var\n",
    "                bn.running_var.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.running_var))\n",
    "                ptr += nb\n",
    "            else:\n",
    "                # Load conv. bias\n",
    "                nb = conv.bias.numel()\n",
    "                conv_b = torch.from_numpy(weights[ptr:ptr + nb]).view_as(conv.bias)\n",
    "                conv.bias.data.copy_(conv_b)\n",
    "                ptr += nb\n",
    "            # Load conv. weights\n",
    "            nw = conv.weight.numel()  # number of weights\n",
    "            conv.weight.data.copy_(torch.from_numpy(weights[ptr:ptr + nw]).view_as(conv.weight))\n",
    "            ptr += nw\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "imgsz = (320, 512)\n",
    "imgsz = 320\n",
    "\n",
    "cfg = 'cfg/yolo-fastest.cfg'\n",
    "weights = 'weights/best.pt'\n",
    "device = 'cpu'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Model Summary: 253 layers, 290066 parameters, 290066 gradients\n"
     ]
    }
   ],
   "source": [
    "# Initialize model\n",
    "model = Darknet(cfg, imgsz)\n",
    "\n",
    "# Load weights\n",
    "if weights.endswith('.pt'):  # pytorch format\n",
    "    model.load_state_dict(torch.load(weights, map_location=device)['model'])\n",
    "else:  # darknet format\n",
    "    load_darknet_weights(model, weights)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "Darknet(\n",
       "  (module_list): ModuleList(\n",
       "    (0): Sequential(\n",
       "      (Conv2d): Conv2d(3, 8, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (1): Sequential(\n",
       "      (Conv2d): Conv2d(8, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (2): Sequential(\n",
       "      (Conv2d): Conv2d(8, 8, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=8, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (3): Sequential(\n",
       "      (Conv2d): Conv2d(8, 4, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(4, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (4): Sequential(\n",
       "      (Conv2d): Conv2d(4, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (5): Sequential(\n",
       "      (Conv2d): Conv2d(8, 8, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=8, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (6): Sequential(\n",
       "      (Conv2d): Conv2d(8, 4, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(4, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (7): Dropout(p=0.15, inplace=False)\n",
       "    (8): WeightedFeatureFusion()\n",
       "    (9): Sequential(\n",
       "      (Conv2d): Conv2d(4, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (10): Sequential(\n",
       "      (Conv2d): Conv2d(24, 24, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=24, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (11): Sequential(\n",
       "      (Conv2d): Conv2d(24, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (12): Sequential(\n",
       "      (Conv2d): Conv2d(8, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(32, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (13): Sequential(\n",
       "      (Conv2d): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=32, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(32, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (14): Sequential(\n",
       "      (Conv2d): Conv2d(32, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (15): Dropout(p=0.15, inplace=False)\n",
       "    (16): WeightedFeatureFusion()\n",
       "    (17): Sequential(\n",
       "      (Conv2d): Conv2d(8, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(32, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (18): Sequential(\n",
       "      (Conv2d): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=32, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(32, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (19): Sequential(\n",
       "      (Conv2d): Conv2d(32, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (20): Dropout(p=0.15, inplace=False)\n",
       "    (21): WeightedFeatureFusion()\n",
       "    (22): Sequential(\n",
       "      (Conv2d): Conv2d(8, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(32, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (23): Sequential(\n",
       "      (Conv2d): Conv2d(32, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=32, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(32, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (24): Sequential(\n",
       "      (Conv2d): Conv2d(32, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (25): Sequential(\n",
       "      (Conv2d): Conv2d(8, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (26): Sequential(\n",
       "      (Conv2d): Conv2d(48, 48, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=48, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (27): Sequential(\n",
       "      (Conv2d): Conv2d(48, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (28): Dropout(p=0.15, inplace=False)\n",
       "    (29): WeightedFeatureFusion()\n",
       "    (30): Sequential(\n",
       "      (Conv2d): Conv2d(8, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (31): Sequential(\n",
       "      (Conv2d): Conv2d(48, 48, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=48, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (32): Sequential(\n",
       "      (Conv2d): Conv2d(48, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(8, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (33): Dropout(p=0.15, inplace=False)\n",
       "    (34): WeightedFeatureFusion()\n",
       "    (35): Sequential(\n",
       "      (Conv2d): Conv2d(8, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (36): Sequential(\n",
       "      (Conv2d): Conv2d(48, 48, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=48, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (37): Sequential(\n",
       "      (Conv2d): Conv2d(48, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(16, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (38): Sequential(\n",
       "      (Conv2d): Conv2d(16, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (39): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (40): Sequential(\n",
       "      (Conv2d): Conv2d(96, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(16, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (41): Dropout(p=0.15, inplace=False)\n",
       "    (42): WeightedFeatureFusion()\n",
       "    (43): Sequential(\n",
       "      (Conv2d): Conv2d(16, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (44): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (45): Sequential(\n",
       "      (Conv2d): Conv2d(96, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(16, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (46): Dropout(p=0.15, inplace=False)\n",
       "    (47): WeightedFeatureFusion()\n",
       "    (48): Sequential(\n",
       "      (Conv2d): Conv2d(16, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (49): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (50): Sequential(\n",
       "      (Conv2d): Conv2d(96, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(16, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (51): Dropout(p=0.15, inplace=False)\n",
       "    (52): WeightedFeatureFusion()\n",
       "    (53): Sequential(\n",
       "      (Conv2d): Conv2d(16, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (54): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (55): Sequential(\n",
       "      (Conv2d): Conv2d(96, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(16, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (56): Dropout(p=0.15, inplace=False)\n",
       "    (57): WeightedFeatureFusion()\n",
       "    (58): Sequential(\n",
       "      (Conv2d): Conv2d(16, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (59): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (60): Sequential(\n",
       "      (Conv2d): Conv2d(96, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (61): Sequential(\n",
       "      (Conv2d): Conv2d(24, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (62): Sequential(\n",
       "      (Conv2d): Conv2d(136, 136, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=136, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (63): Sequential(\n",
       "      (Conv2d): Conv2d(136, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (64): Dropout(p=0.15, inplace=False)\n",
       "    (65): WeightedFeatureFusion()\n",
       "    (66): Sequential(\n",
       "      (Conv2d): Conv2d(24, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (67): Sequential(\n",
       "      (Conv2d): Conv2d(136, 136, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=136, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (68): Sequential(\n",
       "      (Conv2d): Conv2d(136, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (69): Dropout(p=0.15, inplace=False)\n",
       "    (70): WeightedFeatureFusion()\n",
       "    (71): Sequential(\n",
       "      (Conv2d): Conv2d(24, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (72): Sequential(\n",
       "      (Conv2d): Conv2d(136, 136, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=136, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (73): Sequential(\n",
       "      (Conv2d): Conv2d(136, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (74): Dropout(p=0.15, inplace=False)\n",
       "    (75): WeightedFeatureFusion()\n",
       "    (76): Sequential(\n",
       "      (Conv2d): Conv2d(24, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (77): Sequential(\n",
       "      (Conv2d): Conv2d(136, 136, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=136, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (78): Sequential(\n",
       "      (Conv2d): Conv2d(136, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(24, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (79): Dropout(p=0.15, inplace=False)\n",
       "    (80): WeightedFeatureFusion()\n",
       "    (81): Sequential(\n",
       "      (Conv2d): Conv2d(24, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (82): Sequential(\n",
       "      (Conv2d): Conv2d(136, 136, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=136, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(136, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (83): Sequential(\n",
       "      (Conv2d): Conv2d(136, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (84): Sequential(\n",
       "      (Conv2d): Conv2d(48, 224, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (85): Sequential(\n",
       "      (Conv2d): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=224, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (86): Sequential(\n",
       "      (Conv2d): Conv2d(224, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (87): Dropout(p=0.15, inplace=False)\n",
       "    (88): WeightedFeatureFusion()\n",
       "    (89): Sequential(\n",
       "      (Conv2d): Conv2d(48, 224, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (90): Sequential(\n",
       "      (Conv2d): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=224, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (91): Sequential(\n",
       "      (Conv2d): Conv2d(224, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (92): Dropout(p=0.15, inplace=False)\n",
       "    (93): WeightedFeatureFusion()\n",
       "    (94): Sequential(\n",
       "      (Conv2d): Conv2d(48, 224, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (95): Sequential(\n",
       "      (Conv2d): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=224, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (96): Sequential(\n",
       "      (Conv2d): Conv2d(224, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (97): Dropout(p=0.15, inplace=False)\n",
       "    (98): WeightedFeatureFusion()\n",
       "    (99): Sequential(\n",
       "      (Conv2d): Conv2d(48, 224, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (100): Sequential(\n",
       "      (Conv2d): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=224, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (101): Sequential(\n",
       "      (Conv2d): Conv2d(224, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (102): Dropout(p=0.15, inplace=False)\n",
       "    (103): WeightedFeatureFusion()\n",
       "    (104): Sequential(\n",
       "      (Conv2d): Conv2d(48, 224, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (105): Sequential(\n",
       "      (Conv2d): Conv2d(224, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=224, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(224, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (106): Sequential(\n",
       "      (Conv2d): Conv2d(224, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(48, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (107): Dropout(p=0.15, inplace=False)\n",
       "    (108): WeightedFeatureFusion()\n",
       "    (109): Sequential(\n",
       "      (Conv2d): Conv2d(48, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (110): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (111): Sequential(\n",
       "      (Conv2d): Conv2d(96, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(128, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (112): Sequential(\n",
       "      (Conv2d): Conv2d(128, 128, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=128, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(128, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (113): Sequential(\n",
       "      (Conv2d): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(128, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (114): Sequential(\n",
       "      (Conv2d): Conv2d(128, 21, kernel_size=(1, 1), stride=(1, 1))\n",
       "    )\n",
       "    (115): YOLOLayer()\n",
       "    (116): FeatureConcat()\n",
       "    (117): Upsample(scale_factor=2.0, mode=nearest)\n",
       "    (118): FeatureConcat()\n",
       "    (119): Sequential(\n",
       "      (Conv2d): Conv2d(232, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (120): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (121): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (122): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=96, bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "      (activation): LeakyReLU(negative_slope=0.1, inplace=True)\n",
       "    )\n",
       "    (123): Sequential(\n",
       "      (Conv2d): Conv2d(96, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (BatchNorm2d): BatchNorm2d(96, eps=0.0001, momentum=0.03, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (124): Sequential(\n",
       "      (Conv2d): Conv2d(96, 21, kernel_size=(1, 1), stride=(1, 1))\n",
       "    )\n",
       "    (125): YOLOLayer()\n",
       "  )\n",
       ")"
      ]
     },
     "metadata": {},
     "execution_count": 9
    }
   ],
   "source": [
    "model.eval()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 1\n",
    "# Input to the model\n",
    "x = torch.randn(batch_size, 3, 320, 512, requires_grad=True)\n",
    "torch_out = model(x)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "graph(%images : Float(1:491520, 3:163840, 320:512, 512:1),\n      %module_list.0.Conv2d.weight : Float(8:27, 3:9, 3:3, 3:1),\n      %module_list.0.BatchNorm2d.weight : Float(8:1),\n      %module_list.0.BatchNorm2d.bias : Float(8:1),\n      %module_list.0.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.0.BatchNorm2d.running_var : Float(8:1),\n      %module_list.1.Conv2d.weight : Float(8:8, 8:1, 1:1, 1:1),\n      %module_list.1.BatchNorm2d.weight : Float(8:1),\n      %module_list.1.BatchNorm2d.bias : Float(8:1),\n      %module_list.1.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.1.BatchNorm2d.running_var : Float(8:1),\n      %module_list.2.Conv2d.weight : Float(8:9, 1:9, 3:3, 3:1),\n      %module_list.2.BatchNorm2d.weight : Float(8:1),\n      %module_list.2.BatchNorm2d.bias : Float(8:1),\n      %module_list.2.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.2.BatchNorm2d.running_var : Float(8:1),\n      %module_list.3.Conv2d.weight : Float(4:8, 8:1, 1:1, 1:1),\n      %module_list.3.BatchNorm2d.weight : Float(4:1),\n      %module_list.3.BatchNorm2d.bias : Float(4:1),\n      %module_list.3.BatchNorm2d.running_mean : Float(4:1),\n      %module_list.3.BatchNorm2d.running_var : Float(4:1),\n      %module_list.4.Conv2d.weight : Float(8:4, 4:1, 1:1, 1:1),\n      %module_list.4.BatchNorm2d.weight : Float(8:1),\n      %module_list.4.BatchNorm2d.bias : Float(8:1),\n      %module_list.4.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.4.BatchNorm2d.running_var : Float(8:1),\n      %module_list.5.Conv2d.weight : Float(8:9, 1:9, 3:3, 3:1),\n      %module_list.5.BatchNorm2d.weight : Float(8:1),\n      %module_list.5.BatchNorm2d.bias : Float(8:1),\n      %module_list.5.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.5.BatchNorm2d.running_var : Float(8:1),\n      %module_list.6.Conv2d.weight : Float(4:8, 8:1, 1:1, 1:1),\n      %module_list.6.BatchNorm2d.weight : Float(4:1),\n      %module_list.6.BatchNorm2d.bias : Float(4:1),\n      %module_list.6.BatchNorm2d.running_mean : Float(4:1),\n      %module_list.6.BatchNorm2d.running_var : Float(4:1),\n      %module_list.9.Conv2d.weight : Float(24:4, 4:1, 1:1, 1:1),\n      %module_list.9.BatchNorm2d.weight : Float(24:1),\n      %module_list.9.BatchNorm2d.bias : Float(24:1),\n      %module_list.9.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.9.BatchNorm2d.running_var : Float(24:1),\n      %module_list.10.Conv2d.weight : Float(24:9, 1:9, 3:3, 3:1),\n      %module_list.10.BatchNorm2d.weight : Float(24:1),\n      %module_list.10.BatchNorm2d.bias : Float(24:1),\n      %module_list.10.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.10.BatchNorm2d.running_var : Float(24:1),\n      %module_list.11.Conv2d.weight : Float(8:24, 24:1, 1:1, 1:1),\n      %module_list.11.BatchNorm2d.weight : Float(8:1),\n      %module_list.11.BatchNorm2d.bias : Float(8:1),\n      %module_list.11.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.11.BatchNorm2d.running_var : Float(8:1),\n      %module_list.12.Conv2d.weight : Float(32:8, 8:1, 1:1, 1:1),\n      %module_list.12.BatchNorm2d.weight : Float(32:1),\n      %module_list.12.BatchNorm2d.bias : Float(32:1),\n      %module_list.12.BatchNorm2d.running_mean : Float(32:1),\n      %module_list.12.BatchNorm2d.running_var : Float(32:1),\n      %module_list.13.Conv2d.weight : Float(32:9, 1:9, 3:3, 3:1),\n      %module_list.13.BatchNorm2d.weight : Float(32:1),\n      %module_list.13.BatchNorm2d.bias : Float(32:1),\n      %module_list.13.BatchNorm2d.running_mean : Float(32:1),\n      %module_list.13.BatchNorm2d.running_var : Float(32:1),\n      %module_list.14.Conv2d.weight : Float(8:32, 32:1, 1:1, 1:1),\n      %module_list.14.BatchNorm2d.weight : Float(8:1),\n      %module_list.14.BatchNorm2d.bias : Float(8:1),\n      %module_list.14.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.14.BatchNorm2d.running_var : Float(8:1),\n      %module_list.17.Conv2d.weight : Float(32:8, 8:1, 1:1, 1:1),\n      %module_list.17.BatchNorm2d.weight : Float(32:1),\n      %module_list.17.BatchNorm2d.bias : Float(32:1),\n      %module_list.17.BatchNorm2d.running_mean : Float(32:1),\n      %module_list.17.BatchNorm2d.running_var : Float(32:1),\n      %module_list.18.Conv2d.weight : Float(32:9, 1:9, 3:3, 3:1),\n      %module_list.18.BatchNorm2d.weight : Float(32:1),\n      %module_list.18.BatchNorm2d.bias : Float(32:1),\n      %module_list.18.BatchNorm2d.running_mean : Float(32:1),\n      %module_list.18.BatchNorm2d.running_var : Float(32:1),\n      %module_list.19.Conv2d.weight : Float(8:32, 32:1, 1:1, 1:1),\n      %module_list.19.BatchNorm2d.weight : Float(8:1),\n      %module_list.19.BatchNorm2d.bias : Float(8:1),\n      %module_list.19.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.19.BatchNorm2d.running_var : Float(8:1),\n      %module_list.22.Conv2d.weight : Float(32:8, 8:1, 1:1, 1:1),\n      %module_list.22.BatchNorm2d.weight : Float(32:1),\n      %module_list.22.BatchNorm2d.bias : Float(32:1),\n      %module_list.22.BatchNorm2d.running_mean : Float(32:1),\n      %module_list.22.BatchNorm2d.running_var : Float(32:1),\n      %module_list.23.Conv2d.weight : Float(32:9, 1:9, 3:3, 3:1),\n      %module_list.23.BatchNorm2d.weight : Float(32:1),\n      %module_list.23.BatchNorm2d.bias : Float(32:1),\n      %module_list.23.BatchNorm2d.running_mean : Float(32:1),\n      %module_list.23.BatchNorm2d.running_var : Float(32:1),\n      %module_list.24.Conv2d.weight : Float(8:32, 32:1, 1:1, 1:1),\n      %module_list.24.BatchNorm2d.weight : Float(8:1),\n      %module_list.24.BatchNorm2d.bias : Float(8:1),\n      %module_list.24.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.24.BatchNorm2d.running_var : Float(8:1),\n      %module_list.25.Conv2d.weight : Float(48:8, 8:1, 1:1, 1:1),\n      %module_list.25.BatchNorm2d.weight : Float(48:1),\n      %module_list.25.BatchNorm2d.bias : Float(48:1),\n      %module_list.25.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.25.BatchNorm2d.running_var : Float(48:1),\n      %module_list.26.Conv2d.weight : Float(48:9, 1:9, 3:3, 3:1),\n      %module_list.26.BatchNorm2d.weight : Float(48:1),\n      %module_list.26.BatchNorm2d.bias : Float(48:1),\n      %module_list.26.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.26.BatchNorm2d.running_var : Float(48:1),\n      %module_list.27.Conv2d.weight : Float(8:48, 48:1, 1:1, 1:1),\n      %module_list.27.BatchNorm2d.weight : Float(8:1),\n      %module_list.27.BatchNorm2d.bias : Float(8:1),\n      %module_list.27.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.27.BatchNorm2d.running_var : Float(8:1),\n      %module_list.30.Conv2d.weight : Float(48:8, 8:1, 1:1, 1:1),\n      %module_list.30.BatchNorm2d.weight : Float(48:1),\n      %module_list.30.BatchNorm2d.bias : Float(48:1),\n      %module_list.30.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.30.BatchNorm2d.running_var : Float(48:1),\n      %module_list.31.Conv2d.weight : Float(48:9, 1:9, 3:3, 3:1),\n      %module_list.31.BatchNorm2d.weight : Float(48:1),\n      %module_list.31.BatchNorm2d.bias : Float(48:1),\n      %module_list.31.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.31.BatchNorm2d.running_var : Float(48:1),\n      %module_list.32.Conv2d.weight : Float(8:48, 48:1, 1:1, 1:1),\n      %module_list.32.BatchNorm2d.weight : Float(8:1),\n      %module_list.32.BatchNorm2d.bias : Float(8:1),\n      %module_list.32.BatchNorm2d.running_mean : Float(8:1),\n      %module_list.32.BatchNorm2d.running_var : Float(8:1),\n      %module_list.35.Conv2d.weight : Float(48:8, 8:1, 1:1, 1:1),\n      %module_list.35.BatchNorm2d.weight : Float(48:1),\n      %module_list.35.BatchNorm2d.bias : Float(48:1),\n      %module_list.35.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.35.BatchNorm2d.running_var : Float(48:1),\n      %module_list.36.Conv2d.weight : Float(48:9, 1:9, 3:3, 3:1),\n      %module_list.36.BatchNorm2d.weight : Float(48:1),\n      %module_list.36.BatchNorm2d.bias : Float(48:1),\n      %module_list.36.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.36.BatchNorm2d.running_var : Float(48:1),\n      %module_list.37.Conv2d.weight : Float(16:48, 48:1, 1:1, 1:1),\n      %module_list.37.BatchNorm2d.weight : Float(16:1),\n      %module_list.37.BatchNorm2d.bias : Float(16:1),\n      %module_list.37.BatchNorm2d.running_mean : Float(16:1),\n      %module_list.37.BatchNorm2d.running_var : Float(16:1),\n      %module_list.38.Conv2d.weight : Float(96:16, 16:1, 1:1, 1:1),\n      %module_list.38.BatchNorm2d.weight : Float(96:1),\n      %module_list.38.BatchNorm2d.bias : Float(96:1),\n      %module_list.38.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.38.BatchNorm2d.running_var : Float(96:1),\n      %module_list.39.Conv2d.weight : Float(96:9, 1:9, 3:3, 3:1),\n      %module_list.39.BatchNorm2d.weight : Float(96:1),\n      %module_list.39.BatchNorm2d.bias : Float(96:1),\n      %module_list.39.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.39.BatchNorm2d.running_var : Float(96:1),\n      %module_list.40.Conv2d.weight : Float(16:96, 96:1, 1:1, 1:1),\n      %module_list.40.BatchNorm2d.weight : Float(16:1),\n      %module_list.40.BatchNorm2d.bias : Float(16:1),\n      %module_list.40.BatchNorm2d.running_mean : Float(16:1),\n      %module_list.40.BatchNorm2d.running_var : Float(16:1),\n      %module_list.43.Conv2d.weight : Float(96:16, 16:1, 1:1, 1:1),\n      %module_list.43.BatchNorm2d.weight : Float(96:1),\n      %module_list.43.BatchNorm2d.bias : Float(96:1),\n      %module_list.43.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.43.BatchNorm2d.running_var : Float(96:1),\n      %module_list.44.Conv2d.weight : Float(96:9, 1:9, 3:3, 3:1),\n      %module_list.44.BatchNorm2d.weight : Float(96:1),\n      %module_list.44.BatchNorm2d.bias : Float(96:1),\n      %module_list.44.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.44.BatchNorm2d.running_var : Float(96:1),\n      %module_list.45.Conv2d.weight : Float(16:96, 96:1, 1:1, 1:1),\n      %module_list.45.BatchNorm2d.weight : Float(16:1),\n      %module_list.45.BatchNorm2d.bias : Float(16:1),\n      %module_list.45.BatchNorm2d.running_mean : Float(16:1),\n      %module_list.45.BatchNorm2d.running_var : Float(16:1),\n      %module_list.48.Conv2d.weight : Float(96:16, 16:1, 1:1, 1:1),\n      %module_list.48.BatchNorm2d.weight : Float(96:1),\n      %module_list.48.BatchNorm2d.bias : Float(96:1),\n      %module_list.48.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.48.BatchNorm2d.running_var : Float(96:1),\n      %module_list.49.Conv2d.weight : Float(96:9, 1:9, 3:3, 3:1),\n      %module_list.49.BatchNorm2d.weight : Float(96:1),\n      %module_list.49.BatchNorm2d.bias : Float(96:1),\n      %module_list.49.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.49.BatchNorm2d.running_var : Float(96:1),\n      %module_list.50.Conv2d.weight : Float(16:96, 96:1, 1:1, 1:1),\n      %module_list.50.BatchNorm2d.weight : Float(16:1),\n      %module_list.50.BatchNorm2d.bias : Float(16:1),\n      %module_list.50.BatchNorm2d.running_mean : Float(16:1),\n      %module_list.50.BatchNorm2d.running_var : Float(16:1),\n      %module_list.53.Conv2d.weight : Float(96:16, 16:1, 1:1, 1:1),\n      %module_list.53.BatchNorm2d.weight : Float(96:1),\n      %module_list.53.BatchNorm2d.bias : Float(96:1),\n      %module_list.53.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.53.BatchNorm2d.running_var : Float(96:1),\n      %module_list.54.Conv2d.weight : Float(96:9, 1:9, 3:3, 3:1),\n      %module_list.54.BatchNorm2d.weight : Float(96:1),\n      %module_list.54.BatchNorm2d.bias : Float(96:1),\n      %module_list.54.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.54.BatchNorm2d.running_var : Float(96:1),\n      %module_list.55.Conv2d.weight : Float(16:96, 96:1, 1:1, 1:1),\n      %module_list.55.BatchNorm2d.weight : Float(16:1),\n      %module_list.55.BatchNorm2d.bias : Float(16:1),\n      %module_list.55.BatchNorm2d.running_mean : Float(16:1),\n      %module_list.55.BatchNorm2d.running_var : Float(16:1),\n      %module_list.58.Conv2d.weight : Float(96:16, 16:1, 1:1, 1:1),\n      %module_list.58.BatchNorm2d.weight : Float(96:1),\n      %module_list.58.BatchNorm2d.bias : Float(96:1),\n      %module_list.58.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.58.BatchNorm2d.running_var : Float(96:1),\n      %module_list.59.Conv2d.weight : Float(96:9, 1:9, 3:3, 3:1),\n      %module_list.59.BatchNorm2d.weight : Float(96:1),\n      %module_list.59.BatchNorm2d.bias : Float(96:1),\n      %module_list.59.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.59.BatchNorm2d.running_var : Float(96:1),\n      %module_list.60.Conv2d.weight : Float(24:96, 96:1, 1:1, 1:1),\n      %module_list.60.BatchNorm2d.weight : Float(24:1),\n      %module_list.60.BatchNorm2d.bias : Float(24:1),\n      %module_list.60.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.60.BatchNorm2d.running_var : Float(24:1),\n      %module_list.61.Conv2d.weight : Float(136:24, 24:1, 1:1, 1:1),\n      %module_list.61.BatchNorm2d.weight : Float(136:1),\n      %module_list.61.BatchNorm2d.bias : Float(136:1),\n      %module_list.61.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.61.BatchNorm2d.running_var : Float(136:1),\n      %module_list.62.Conv2d.weight : Float(136:9, 1:9, 3:3, 3:1),\n      %module_list.62.BatchNorm2d.weight : Float(136:1),\n      %module_list.62.BatchNorm2d.bias : Float(136:1),\n      %module_list.62.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.62.BatchNorm2d.running_var : Float(136:1),\n      %module_list.63.Conv2d.weight : Float(24:136, 136:1, 1:1, 1:1),\n      %module_list.63.BatchNorm2d.weight : Float(24:1),\n      %module_list.63.BatchNorm2d.bias : Float(24:1),\n      %module_list.63.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.63.BatchNorm2d.running_var : Float(24:1),\n      %module_list.66.Conv2d.weight : Float(136:24, 24:1, 1:1, 1:1),\n      %module_list.66.BatchNorm2d.weight : Float(136:1),\n      %module_list.66.BatchNorm2d.bias : Float(136:1),\n      %module_list.66.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.66.BatchNorm2d.running_var : Float(136:1),\n      %module_list.67.Conv2d.weight : Float(136:9, 1:9, 3:3, 3:1),\n      %module_list.67.BatchNorm2d.weight : Float(136:1),\n      %module_list.67.BatchNorm2d.bias : Float(136:1),\n      %module_list.67.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.67.BatchNorm2d.running_var : Float(136:1),\n      %module_list.68.Conv2d.weight : Float(24:136, 136:1, 1:1, 1:1),\n      %module_list.68.BatchNorm2d.weight : Float(24:1),\n      %module_list.68.BatchNorm2d.bias : Float(24:1),\n      %module_list.68.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.68.BatchNorm2d.running_var : Float(24:1),\n      %module_list.71.Conv2d.weight : Float(136:24, 24:1, 1:1, 1:1),\n      %module_list.71.BatchNorm2d.weight : Float(136:1),\n      %module_list.71.BatchNorm2d.bias : Float(136:1),\n      %module_list.71.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.71.BatchNorm2d.running_var : Float(136:1),\n      %module_list.72.Conv2d.weight : Float(136:9, 1:9, 3:3, 3:1),\n      %module_list.72.BatchNorm2d.weight : Float(136:1),\n      %module_list.72.BatchNorm2d.bias : Float(136:1),\n      %module_list.72.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.72.BatchNorm2d.running_var : Float(136:1),\n      %module_list.73.Conv2d.weight : Float(24:136, 136:1, 1:1, 1:1),\n      %module_list.73.BatchNorm2d.weight : Float(24:1),\n      %module_list.73.BatchNorm2d.bias : Float(24:1),\n      %module_list.73.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.73.BatchNorm2d.running_var : Float(24:1),\n      %module_list.76.Conv2d.weight : Float(136:24, 24:1, 1:1, 1:1),\n      %module_list.76.BatchNorm2d.weight : Float(136:1),\n      %module_list.76.BatchNorm2d.bias : Float(136:1),\n      %module_list.76.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.76.BatchNorm2d.running_var : Float(136:1),\n      %module_list.77.Conv2d.weight : Float(136:9, 1:9, 3:3, 3:1),\n      %module_list.77.BatchNorm2d.weight : Float(136:1),\n      %module_list.77.BatchNorm2d.bias : Float(136:1),\n      %module_list.77.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.77.BatchNorm2d.running_var : Float(136:1),\n      %module_list.78.Conv2d.weight : Float(24:136, 136:1, 1:1, 1:1),\n      %module_list.78.BatchNorm2d.weight : Float(24:1),\n      %module_list.78.BatchNorm2d.bias : Float(24:1),\n      %module_list.78.BatchNorm2d.running_mean : Float(24:1),\n      %module_list.78.BatchNorm2d.running_var : Float(24:1),\n      %module_list.81.Conv2d.weight : Float(136:24, 24:1, 1:1, 1:1),\n      %module_list.81.BatchNorm2d.weight : Float(136:1),\n      %module_list.81.BatchNorm2d.bias : Float(136:1),\n      %module_list.81.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.81.BatchNorm2d.running_var : Float(136:1),\n      %module_list.82.Conv2d.weight : Float(136:9, 1:9, 3:3, 3:1),\n      %module_list.82.BatchNorm2d.weight : Float(136:1),\n      %module_list.82.BatchNorm2d.bias : Float(136:1),\n      %module_list.82.BatchNorm2d.running_mean : Float(136:1),\n      %module_list.82.BatchNorm2d.running_var : Float(136:1),\n      %module_list.83.Conv2d.weight : Float(48:136, 136:1, 1:1, 1:1),\n      %module_list.83.BatchNorm2d.weight : Float(48:1),\n      %module_list.83.BatchNorm2d.bias : Float(48:1),\n      %module_list.83.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.83.BatchNorm2d.running_var : Float(48:1),\n      %module_list.84.Conv2d.weight : Float(224:48, 48:1, 1:1, 1:1),\n      %module_list.84.BatchNorm2d.weight : Float(224:1),\n      %module_list.84.BatchNorm2d.bias : Float(224:1),\n      %module_list.84.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.84.BatchNorm2d.running_var : Float(224:1),\n      %module_list.85.Conv2d.weight : Float(224:9, 1:9, 3:3, 3:1),\n      %module_list.85.BatchNorm2d.weight : Float(224:1),\n      %module_list.85.BatchNorm2d.bias : Float(224:1),\n      %module_list.85.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.85.BatchNorm2d.running_var : Float(224:1),\n      %module_list.86.Conv2d.weight : Float(48:224, 224:1, 1:1, 1:1),\n      %module_list.86.BatchNorm2d.weight : Float(48:1),\n      %module_list.86.BatchNorm2d.bias : Float(48:1),\n      %module_list.86.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.86.BatchNorm2d.running_var : Float(48:1),\n      %module_list.89.Conv2d.weight : Float(224:48, 48:1, 1:1, 1:1),\n      %module_list.89.BatchNorm2d.weight : Float(224:1),\n      %module_list.89.BatchNorm2d.bias : Float(224:1),\n      %module_list.89.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.89.BatchNorm2d.running_var : Float(224:1),\n      %module_list.90.Conv2d.weight : Float(224:9, 1:9, 3:3, 3:1),\n      %module_list.90.BatchNorm2d.weight : Float(224:1),\n      %module_list.90.BatchNorm2d.bias : Float(224:1),\n      %module_list.90.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.90.BatchNorm2d.running_var : Float(224:1),\n      %module_list.91.Conv2d.weight : Float(48:224, 224:1, 1:1, 1:1),\n      %module_list.91.BatchNorm2d.weight : Float(48:1),\n      %module_list.91.BatchNorm2d.bias : Float(48:1),\n      %module_list.91.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.91.BatchNorm2d.running_var : Float(48:1),\n      %module_list.94.Conv2d.weight : Float(224:48, 48:1, 1:1, 1:1),\n      %module_list.94.BatchNorm2d.weight : Float(224:1),\n      %module_list.94.BatchNorm2d.bias : Float(224:1),\n      %module_list.94.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.94.BatchNorm2d.running_var : Float(224:1),\n      %module_list.95.Conv2d.weight : Float(224:9, 1:9, 3:3, 3:1),\n      %module_list.95.BatchNorm2d.weight : Float(224:1),\n      %module_list.95.BatchNorm2d.bias : Float(224:1),\n      %module_list.95.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.95.BatchNorm2d.running_var : Float(224:1),\n      %module_list.96.Conv2d.weight : Float(48:224, 224:1, 1:1, 1:1),\n      %module_list.96.BatchNorm2d.weight : Float(48:1),\n      %module_list.96.BatchNorm2d.bias : Float(48:1),\n      %module_list.96.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.96.BatchNorm2d.running_var : Float(48:1),\n      %module_list.99.Conv2d.weight : Float(224:48, 48:1, 1:1, 1:1),\n      %module_list.99.BatchNorm2d.weight : Float(224:1),\n      %module_list.99.BatchNorm2d.bias : Float(224:1),\n      %module_list.99.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.99.BatchNorm2d.running_var : Float(224:1),\n      %module_list.100.Conv2d.weight : Float(224:9, 1:9, 3:3, 3:1),\n      %module_list.100.BatchNorm2d.weight : Float(224:1),\n      %module_list.100.BatchNorm2d.bias : Float(224:1),\n      %module_list.100.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.100.BatchNorm2d.running_var : Float(224:1),\n      %module_list.101.Conv2d.weight : Float(48:224, 224:1, 1:1, 1:1),\n      %module_list.101.BatchNorm2d.weight : Float(48:1),\n      %module_list.101.BatchNorm2d.bias : Float(48:1),\n      %module_list.101.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.101.BatchNorm2d.running_var : Float(48:1),\n      %module_list.104.Conv2d.weight : Float(224:48, 48:1, 1:1, 1:1),\n      %module_list.104.BatchNorm2d.weight : Float(224:1),\n      %module_list.104.BatchNorm2d.bias : Float(224:1),\n      %module_list.104.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.104.BatchNorm2d.running_var : Float(224:1),\n      %module_list.105.Conv2d.weight : Float(224:9, 1:9, 3:3, 3:1),\n      %module_list.105.BatchNorm2d.weight : Float(224:1),\n      %module_list.105.BatchNorm2d.bias : Float(224:1),\n      %module_list.105.BatchNorm2d.running_mean : Float(224:1),\n      %module_list.105.BatchNorm2d.running_var : Float(224:1),\n      %module_list.106.Conv2d.weight : Float(48:224, 224:1, 1:1, 1:1),\n      %module_list.106.BatchNorm2d.weight : Float(48:1),\n      %module_list.106.BatchNorm2d.bias : Float(48:1),\n      %module_list.106.BatchNorm2d.running_mean : Float(48:1),\n      %module_list.106.BatchNorm2d.running_var : Float(48:1),\n      %module_list.109.Conv2d.weight : Float(96:48, 48:1, 1:1, 1:1),\n      %module_list.109.BatchNorm2d.weight : Float(96:1),\n      %module_list.109.BatchNorm2d.bias : Float(96:1),\n      %module_list.109.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.109.BatchNorm2d.running_var : Float(96:1),\n      %module_list.110.Conv2d.weight : Float(96:25, 1:25, 5:5, 5:1),\n      %module_list.110.BatchNorm2d.weight : Float(96:1),\n      %module_list.110.BatchNorm2d.bias : Float(96:1),\n      %module_list.110.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.110.BatchNorm2d.running_var : Float(96:1),\n      %module_list.111.Conv2d.weight : Float(128:96, 96:1, 1:1, 1:1),\n      %module_list.111.BatchNorm2d.weight : Float(128:1),\n      %module_list.111.BatchNorm2d.bias : Float(128:1),\n      %module_list.111.BatchNorm2d.running_mean : Float(128:1),\n      %module_list.111.BatchNorm2d.running_var : Float(128:1),\n      %module_list.112.Conv2d.weight : Float(128:25, 1:25, 5:5, 5:1),\n      %module_list.112.BatchNorm2d.weight : Float(128:1),\n      %module_list.112.BatchNorm2d.bias : Float(128:1),\n      %module_list.112.BatchNorm2d.running_mean : Float(128:1),\n      %module_list.112.BatchNorm2d.running_var : Float(128:1),\n      %module_list.113.Conv2d.weight : Float(128:128, 128:1, 1:1, 1:1),\n      %module_list.113.BatchNorm2d.weight : Float(128:1),\n      %module_list.113.BatchNorm2d.bias : Float(128:1),\n      %module_list.113.BatchNorm2d.running_mean : Float(128:1),\n      %module_list.113.BatchNorm2d.running_var : Float(128:1),\n      %module_list.114.Conv2d.weight : Float(21:128, 128:1, 1:1, 1:1),\n      %module_list.114.Conv2d.bias : Float(21:1),\n      %module_list.119.Conv2d.weight : Float(96:232, 232:1, 1:1, 1:1),\n      %module_list.119.BatchNorm2d.weight : Float(96:1),\n      %module_list.119.BatchNorm2d.bias : Float(96:1),\n      %module_list.119.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.119.BatchNorm2d.running_var : Float(96:1),\n      %module_list.120.Conv2d.weight : Float(96:25, 1:25, 5:5, 5:1),\n      %module_list.120.BatchNorm2d.weight : Float(96:1),\n      %module_list.120.BatchNorm2d.bias : Float(96:1),\n      %module_list.120.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.120.BatchNorm2d.running_var : Float(96:1),\n      %module_list.121.Conv2d.weight : Float(96:96, 96:1, 1:1, 1:1),\n      %module_list.121.BatchNorm2d.weight : Float(96:1),\n      %module_list.121.BatchNorm2d.bias : Float(96:1),\n      %module_list.121.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.121.BatchNorm2d.running_var : Float(96:1),\n      %module_list.122.Conv2d.weight : Float(96:25, 1:25, 5:5, 5:1),\n      %module_list.122.BatchNorm2d.weight : Float(96:1),\n      %module_list.122.BatchNorm2d.bias : Float(96:1),\n      %module_list.122.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.122.BatchNorm2d.running_var : Float(96:1),\n      %module_list.123.Conv2d.weight : Float(96:96, 96:1, 1:1, 1:1),\n      %module_list.123.BatchNorm2d.weight : Float(96:1),\n      %module_list.123.BatchNorm2d.bias : Float(96:1),\n      %module_list.123.BatchNorm2d.running_mean : Float(96:1),\n      %module_list.123.BatchNorm2d.running_var : Float(96:1),\n      %module_list.124.Conv2d.weight : Float(21:96, 96:1, 1:1, 1:1),\n      %module_list.124.Conv2d.bias : Float(21:1),\n      %1536 : Long(1:1),\n      %1537 : Long(1:1),\n      %1538 : Long(1:1),\n      %1539 : Long(1:1),\n      %1540 : Long(1:1),\n      %1541 : Long(1:1),\n      %1542 : Long(1:1),\n      %1543 : Long(1:1),\n      %1544 : Long(1:1),\n      %1549 : Float(4:1),\n      %1550 : Long(1:1),\n      %1551 : Long(1:1),\n      %1552 : Long(1:1),\n      %1553 : Long(1:1),\n      %1554 : Long(1:1),\n      %1555 : Long(1:1),\n      %1556 : Long(1:1),\n      %1557 : Long(1:1),\n      %1558 : Long(1:1)):\n  %503 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%images, %module_list.0.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %504 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%503, %module_list.0.BatchNorm2d.weight, %module_list.0.BatchNorm2d.bias, %module_list.0.BatchNorm2d.running_mean, %module_list.0.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %505 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%504) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %506 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%505, %module_list.1.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %507 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%506, %module_list.1.BatchNorm2d.weight, %module_list.1.BatchNorm2d.bias, %module_list.1.BatchNorm2d.running_mean, %module_list.1.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %508 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%507) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %509 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=8, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%508, %module_list.2.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %510 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%509, %module_list.2.BatchNorm2d.weight, %module_list.2.BatchNorm2d.bias, %module_list.2.BatchNorm2d.running_mean, %module_list.2.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %511 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%510) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %512 : Float(1:163840, 4:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%511, %module_list.3.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %513 : Float(1:163840, 4:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%512, %module_list.3.BatchNorm2d.weight, %module_list.3.BatchNorm2d.bias, %module_list.3.BatchNorm2d.running_mean, %module_list.3.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %514 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%513, %module_list.4.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %515 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%514, %module_list.4.BatchNorm2d.weight, %module_list.4.BatchNorm2d.bias, %module_list.4.BatchNorm2d.running_mean, %module_list.4.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %516 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%515) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %517 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=8, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%516, %module_list.5.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %518 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%517, %module_list.5.BatchNorm2d.weight, %module_list.5.BatchNorm2d.bias, %module_list.5.BatchNorm2d.running_mean, %module_list.5.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %519 : Float(1:327680, 8:40960, 160:256, 256:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%518) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %520 : Float(1:163840, 4:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%519, %module_list.6.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %521 : Float(1:163840, 4:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%520, %module_list.6.BatchNorm2d.weight, %module_list.6.BatchNorm2d.bias, %module_list.6.BatchNorm2d.running_mean, %module_list.6.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %522 : Float(1:163840, 4:40960, 160:256, 256:1) = onnx::Add(%521, %513) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %523 : Float(1:983040, 24:40960, 160:256, 256:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%522, %module_list.9.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %524 : Float(1:983040, 24:40960, 160:256, 256:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%523, %module_list.9.BatchNorm2d.weight, %module_list.9.BatchNorm2d.bias, %module_list.9.BatchNorm2d.running_mean, %module_list.9.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %525 : Float(1:983040, 24:40960, 160:256, 256:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%524) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %526 : Float(1:245760, 24:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=24, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%525, %module_list.10.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %527 : Float(1:245760, 24:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%526, %module_list.10.BatchNorm2d.weight, %module_list.10.BatchNorm2d.bias, %module_list.10.BatchNorm2d.running_mean, %module_list.10.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %528 : Float(1:245760, 24:10240, 80:128, 128:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%527) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %529 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%528, %module_list.11.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %530 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%529, %module_list.11.BatchNorm2d.weight, %module_list.11.BatchNorm2d.bias, %module_list.11.BatchNorm2d.running_mean, %module_list.11.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %531 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%530, %module_list.12.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %532 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%531, %module_list.12.BatchNorm2d.weight, %module_list.12.BatchNorm2d.bias, %module_list.12.BatchNorm2d.running_mean, %module_list.12.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %533 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%532) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %534 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=32, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%533, %module_list.13.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %535 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%534, %module_list.13.BatchNorm2d.weight, %module_list.13.BatchNorm2d.bias, %module_list.13.BatchNorm2d.running_mean, %module_list.13.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %536 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%535) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %537 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%536, %module_list.14.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %538 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%537, %module_list.14.BatchNorm2d.weight, %module_list.14.BatchNorm2d.bias, %module_list.14.BatchNorm2d.running_mean, %module_list.14.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %539 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::Add(%538, %530) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %540 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%539, %module_list.17.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %541 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%540, %module_list.17.BatchNorm2d.weight, %module_list.17.BatchNorm2d.bias, %module_list.17.BatchNorm2d.running_mean, %module_list.17.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %542 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%541) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %543 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=32, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%542, %module_list.18.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %544 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%543, %module_list.18.BatchNorm2d.weight, %module_list.18.BatchNorm2d.bias, %module_list.18.BatchNorm2d.running_mean, %module_list.18.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %545 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%544) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %546 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%545, %module_list.19.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %547 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%546, %module_list.19.BatchNorm2d.weight, %module_list.19.BatchNorm2d.bias, %module_list.19.BatchNorm2d.running_mean, %module_list.19.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %548 : Float(1:81920, 8:10240, 80:128, 128:1) = onnx::Add(%547, %539) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %549 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%548, %module_list.22.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %550 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%549, %module_list.22.BatchNorm2d.weight, %module_list.22.BatchNorm2d.bias, %module_list.22.BatchNorm2d.running_mean, %module_list.22.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %551 : Float(1:327680, 32:10240, 80:128, 128:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%550) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %552 : Float(1:81920, 32:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=32, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%551, %module_list.23.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %553 : Float(1:81920, 32:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%552, %module_list.23.BatchNorm2d.weight, %module_list.23.BatchNorm2d.bias, %module_list.23.BatchNorm2d.running_mean, %module_list.23.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %554 : Float(1:81920, 32:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%553) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %555 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%554, %module_list.24.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %556 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%555, %module_list.24.BatchNorm2d.weight, %module_list.24.BatchNorm2d.bias, %module_list.24.BatchNorm2d.running_mean, %module_list.24.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %557 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%556, %module_list.25.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %558 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%557, %module_list.25.BatchNorm2d.weight, %module_list.25.BatchNorm2d.bias, %module_list.25.BatchNorm2d.running_mean, %module_list.25.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %559 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%558) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %560 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=48, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%559, %module_list.26.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %561 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%560, %module_list.26.BatchNorm2d.weight, %module_list.26.BatchNorm2d.bias, %module_list.26.BatchNorm2d.running_mean, %module_list.26.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %562 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%561) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %563 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%562, %module_list.27.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %564 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%563, %module_list.27.BatchNorm2d.weight, %module_list.27.BatchNorm2d.bias, %module_list.27.BatchNorm2d.running_mean, %module_list.27.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %565 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::Add(%564, %556) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %566 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%565, %module_list.30.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %567 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%566, %module_list.30.BatchNorm2d.weight, %module_list.30.BatchNorm2d.bias, %module_list.30.BatchNorm2d.running_mean, %module_list.30.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %568 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%567) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %569 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=48, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%568, %module_list.31.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %570 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%569, %module_list.31.BatchNorm2d.weight, %module_list.31.BatchNorm2d.bias, %module_list.31.BatchNorm2d.running_mean, %module_list.31.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %571 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%570) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %572 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%571, %module_list.32.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %573 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%572, %module_list.32.BatchNorm2d.weight, %module_list.32.BatchNorm2d.bias, %module_list.32.BatchNorm2d.running_mean, %module_list.32.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %574 : Float(1:20480, 8:2560, 40:64, 64:1) = onnx::Add(%573, %565) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %575 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%574, %module_list.35.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %576 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%575, %module_list.35.BatchNorm2d.weight, %module_list.35.BatchNorm2d.bias, %module_list.35.BatchNorm2d.running_mean, %module_list.35.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %577 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%576) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %578 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=48, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%577, %module_list.36.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %579 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%578, %module_list.36.BatchNorm2d.weight, %module_list.36.BatchNorm2d.bias, %module_list.36.BatchNorm2d.running_mean, %module_list.36.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %580 : Float(1:122880, 48:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%579) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %581 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%580, %module_list.37.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %582 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%581, %module_list.37.BatchNorm2d.weight, %module_list.37.BatchNorm2d.bias, %module_list.37.BatchNorm2d.running_mean, %module_list.37.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %583 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%582, %module_list.38.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %584 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%583, %module_list.38.BatchNorm2d.weight, %module_list.38.BatchNorm2d.bias, %module_list.38.BatchNorm2d.running_mean, %module_list.38.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %585 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%584) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %586 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%585, %module_list.39.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %587 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%586, %module_list.39.BatchNorm2d.weight, %module_list.39.BatchNorm2d.bias, %module_list.39.BatchNorm2d.running_mean, %module_list.39.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %588 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%587) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %589 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%588, %module_list.40.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %590 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%589, %module_list.40.BatchNorm2d.weight, %module_list.40.BatchNorm2d.bias, %module_list.40.BatchNorm2d.running_mean, %module_list.40.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %591 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Add(%590, %582) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %592 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%591, %module_list.43.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %593 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%592, %module_list.43.BatchNorm2d.weight, %module_list.43.BatchNorm2d.bias, %module_list.43.BatchNorm2d.running_mean, %module_list.43.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %594 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%593) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %595 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%594, %module_list.44.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %596 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%595, %module_list.44.BatchNorm2d.weight, %module_list.44.BatchNorm2d.bias, %module_list.44.BatchNorm2d.running_mean, %module_list.44.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %597 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%596) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %598 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%597, %module_list.45.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %599 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%598, %module_list.45.BatchNorm2d.weight, %module_list.45.BatchNorm2d.bias, %module_list.45.BatchNorm2d.running_mean, %module_list.45.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %600 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Add(%599, %591) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %601 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%600, %module_list.48.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %602 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%601, %module_list.48.BatchNorm2d.weight, %module_list.48.BatchNorm2d.bias, %module_list.48.BatchNorm2d.running_mean, %module_list.48.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %603 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%602) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %604 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%603, %module_list.49.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %605 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%604, %module_list.49.BatchNorm2d.weight, %module_list.49.BatchNorm2d.bias, %module_list.49.BatchNorm2d.running_mean, %module_list.49.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %606 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%605) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %607 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%606, %module_list.50.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %608 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%607, %module_list.50.BatchNorm2d.weight, %module_list.50.BatchNorm2d.bias, %module_list.50.BatchNorm2d.running_mean, %module_list.50.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %609 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Add(%608, %600) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %610 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%609, %module_list.53.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %611 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%610, %module_list.53.BatchNorm2d.weight, %module_list.53.BatchNorm2d.bias, %module_list.53.BatchNorm2d.running_mean, %module_list.53.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %612 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%611) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %613 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%612, %module_list.54.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %614 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%613, %module_list.54.BatchNorm2d.weight, %module_list.54.BatchNorm2d.bias, %module_list.54.BatchNorm2d.running_mean, %module_list.54.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %615 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%614) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %616 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%615, %module_list.55.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %617 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%616, %module_list.55.BatchNorm2d.weight, %module_list.55.BatchNorm2d.bias, %module_list.55.BatchNorm2d.running_mean, %module_list.55.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %618 : Float(1:40960, 16:2560, 40:64, 64:1) = onnx::Add(%617, %609) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %619 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%618, %module_list.58.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %620 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%619, %module_list.58.BatchNorm2d.weight, %module_list.58.BatchNorm2d.bias, %module_list.58.BatchNorm2d.running_mean, %module_list.58.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %621 : Float(1:245760, 96:2560, 40:64, 64:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%620) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %622 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%621, %module_list.59.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %623 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%622, %module_list.59.BatchNorm2d.weight, %module_list.59.BatchNorm2d.bias, %module_list.59.BatchNorm2d.running_mean, %module_list.59.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %624 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%623) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %625 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%624, %module_list.60.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %626 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%625, %module_list.60.BatchNorm2d.weight, %module_list.60.BatchNorm2d.bias, %module_list.60.BatchNorm2d.running_mean, %module_list.60.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %627 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%626, %module_list.61.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %628 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%627, %module_list.61.BatchNorm2d.weight, %module_list.61.BatchNorm2d.bias, %module_list.61.BatchNorm2d.running_mean, %module_list.61.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %629 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%628) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %630 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=136, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%629, %module_list.62.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %631 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%630, %module_list.62.BatchNorm2d.weight, %module_list.62.BatchNorm2d.bias, %module_list.62.BatchNorm2d.running_mean, %module_list.62.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %632 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%631) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %633 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%632, %module_list.63.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %634 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%633, %module_list.63.BatchNorm2d.weight, %module_list.63.BatchNorm2d.bias, %module_list.63.BatchNorm2d.running_mean, %module_list.63.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %635 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Add(%634, %626) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %636 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%635, %module_list.66.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %637 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%636, %module_list.66.BatchNorm2d.weight, %module_list.66.BatchNorm2d.bias, %module_list.66.BatchNorm2d.running_mean, %module_list.66.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %638 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%637) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %639 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=136, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%638, %module_list.67.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %640 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%639, %module_list.67.BatchNorm2d.weight, %module_list.67.BatchNorm2d.bias, %module_list.67.BatchNorm2d.running_mean, %module_list.67.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %641 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%640) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %642 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%641, %module_list.68.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %643 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%642, %module_list.68.BatchNorm2d.weight, %module_list.68.BatchNorm2d.bias, %module_list.68.BatchNorm2d.running_mean, %module_list.68.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %644 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Add(%643, %635) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %645 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%644, %module_list.71.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %646 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%645, %module_list.71.BatchNorm2d.weight, %module_list.71.BatchNorm2d.bias, %module_list.71.BatchNorm2d.running_mean, %module_list.71.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %647 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%646) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %648 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=136, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%647, %module_list.72.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %649 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%648, %module_list.72.BatchNorm2d.weight, %module_list.72.BatchNorm2d.bias, %module_list.72.BatchNorm2d.running_mean, %module_list.72.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %650 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%649) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %651 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%650, %module_list.73.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %652 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%651, %module_list.73.BatchNorm2d.weight, %module_list.73.BatchNorm2d.bias, %module_list.73.BatchNorm2d.running_mean, %module_list.73.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %653 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Add(%652, %644) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %654 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%653, %module_list.76.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %655 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%654, %module_list.76.BatchNorm2d.weight, %module_list.76.BatchNorm2d.bias, %module_list.76.BatchNorm2d.running_mean, %module_list.76.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %656 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%655) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %657 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=136, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%656, %module_list.77.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %658 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%657, %module_list.77.BatchNorm2d.weight, %module_list.77.BatchNorm2d.bias, %module_list.77.BatchNorm2d.running_mean, %module_list.77.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %659 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%658) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %660 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%659, %module_list.78.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %661 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%660, %module_list.78.BatchNorm2d.weight, %module_list.78.BatchNorm2d.bias, %module_list.78.BatchNorm2d.running_mean, %module_list.78.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %662 : Float(1:15360, 24:640, 20:32, 32:1) = onnx::Add(%661, %653) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %663 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%662, %module_list.81.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %664 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%663, %module_list.81.BatchNorm2d.weight, %module_list.81.BatchNorm2d.bias, %module_list.81.BatchNorm2d.running_mean, %module_list.81.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %665 : Float(1:87040, 136:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%664) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %666 : Float(1:21760, 136:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=136, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%665, %module_list.82.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %667 : Float(1:21760, 136:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%666, %module_list.82.BatchNorm2d.weight, %module_list.82.BatchNorm2d.bias, %module_list.82.BatchNorm2d.running_mean, %module_list.82.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %668 : Float(1:21760, 136:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%667) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %669 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%668, %module_list.83.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %670 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%669, %module_list.83.BatchNorm2d.weight, %module_list.83.BatchNorm2d.bias, %module_list.83.BatchNorm2d.running_mean, %module_list.83.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %671 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%670, %module_list.84.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %672 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%671, %module_list.84.BatchNorm2d.weight, %module_list.84.BatchNorm2d.bias, %module_list.84.BatchNorm2d.running_mean, %module_list.84.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %673 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%672) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %674 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=224, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%673, %module_list.85.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %675 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%674, %module_list.85.BatchNorm2d.weight, %module_list.85.BatchNorm2d.bias, %module_list.85.BatchNorm2d.running_mean, %module_list.85.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %676 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%675) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %677 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%676, %module_list.86.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %678 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%677, %module_list.86.BatchNorm2d.weight, %module_list.86.BatchNorm2d.bias, %module_list.86.BatchNorm2d.running_mean, %module_list.86.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %679 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Add(%678, %670) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %680 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%679, %module_list.89.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %681 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%680, %module_list.89.BatchNorm2d.weight, %module_list.89.BatchNorm2d.bias, %module_list.89.BatchNorm2d.running_mean, %module_list.89.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %682 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%681) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %683 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=224, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%682, %module_list.90.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %684 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%683, %module_list.90.BatchNorm2d.weight, %module_list.90.BatchNorm2d.bias, %module_list.90.BatchNorm2d.running_mean, %module_list.90.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %685 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%684) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %686 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%685, %module_list.91.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %687 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%686, %module_list.91.BatchNorm2d.weight, %module_list.91.BatchNorm2d.bias, %module_list.91.BatchNorm2d.running_mean, %module_list.91.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %688 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Add(%687, %679) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %689 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%688, %module_list.94.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %690 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%689, %module_list.94.BatchNorm2d.weight, %module_list.94.BatchNorm2d.bias, %module_list.94.BatchNorm2d.running_mean, %module_list.94.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %691 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%690) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %692 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=224, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%691, %module_list.95.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %693 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%692, %module_list.95.BatchNorm2d.weight, %module_list.95.BatchNorm2d.bias, %module_list.95.BatchNorm2d.running_mean, %module_list.95.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %694 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%693) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %695 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%694, %module_list.96.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %696 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%695, %module_list.96.BatchNorm2d.weight, %module_list.96.BatchNorm2d.bias, %module_list.96.BatchNorm2d.running_mean, %module_list.96.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %697 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Add(%696, %688) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %698 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%697, %module_list.99.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %699 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%698, %module_list.99.BatchNorm2d.weight, %module_list.99.BatchNorm2d.bias, %module_list.99.BatchNorm2d.running_mean, %module_list.99.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %700 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%699) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %701 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=224, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%700, %module_list.100.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %702 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%701, %module_list.100.BatchNorm2d.weight, %module_list.100.BatchNorm2d.bias, %module_list.100.BatchNorm2d.running_mean, %module_list.100.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %703 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%702) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %704 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%703, %module_list.101.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %705 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%704, %module_list.101.BatchNorm2d.weight, %module_list.101.BatchNorm2d.bias, %module_list.101.BatchNorm2d.running_mean, %module_list.101.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %706 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Add(%705, %697) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %707 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%706, %module_list.104.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %708 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%707, %module_list.104.BatchNorm2d.weight, %module_list.104.BatchNorm2d.bias, %module_list.104.BatchNorm2d.running_mean, %module_list.104.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %709 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%708) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %710 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=224, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%709, %module_list.105.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %711 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%710, %module_list.105.BatchNorm2d.weight, %module_list.105.BatchNorm2d.bias, %module_list.105.BatchNorm2d.running_mean, %module_list.105.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %712 : Float(1:35840, 224:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%711) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %713 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%712, %module_list.106.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %714 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%713, %module_list.106.BatchNorm2d.weight, %module_list.106.BatchNorm2d.bias, %module_list.106.BatchNorm2d.running_mean, %module_list.106.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:973:0\n  %715 : Float(1:7680, 48:160, 10:16, 16:1) = onnx::Add(%714, %706) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:61:0\n  %716 : Float(1:15360, 96:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%715, %module_list.109.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %717 : Float(1:15360, 96:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%716, %module_list.109.BatchNorm2d.weight, %module_list.109.BatchNorm2d.bias, %module_list.109.BatchNorm2d.running_mean, %module_list.109.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %718 : Float(1:15360, 96:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%717) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %719 : Float(1:15360, 96:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[5, 5], pads=[2, 2, 2, 2], strides=[1, 1]](%718, %module_list.110.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %720 : Float(1:15360, 96:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%719, %module_list.110.BatchNorm2d.weight, %module_list.110.BatchNorm2d.bias, %module_list.110.BatchNorm2d.running_mean, %module_list.110.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %721 : Float(1:15360, 96:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%720) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %722 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%721, %module_list.111.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %723 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%722, %module_list.111.BatchNorm2d.weight, %module_list.111.BatchNorm2d.bias, %module_list.111.BatchNorm2d.running_mean, %module_list.111.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %724 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=128, kernel_shape=[5, 5], pads=[2, 2, 2, 2], strides=[1, 1]](%723, %module_list.112.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %725 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%724, %module_list.112.BatchNorm2d.weight, %module_list.112.BatchNorm2d.bias, %module_list.112.BatchNorm2d.running_mean, %module_list.112.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %726 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%725) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %727 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%726, %module_list.113.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %728 : Float(1:20480, 128:160, 10:16, 16:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%727, %module_list.113.BatchNorm2d.weight, %module_list.113.BatchNorm2d.bias, %module_list.113.BatchNorm2d.running_mean, %module_list.113.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %729 : Float(1:3360, 21:160, 10:16, 16:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%728, %module_list.114.Conv2d.weight, %module_list.114.Conv2d.bias) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %730 : Tensor = onnx::Shape(%729)\n  %731 : Tensor = onnx::Constant[value={0}]()\n  %732 : Long() = onnx::Gather[axis=0](%730, %731) # <ipython-input-6-cc76a8bf4ad5>:194:0\n  %737 : Tensor = onnx::Unsqueeze[axes=[0]](%732)\n  %742 : Tensor = onnx::Concat[axis=0](%737, %1536, %1537, %1538, %1539)\n  %743 : Float(1:3360, 3:1120, 7:160, 10:16, 16:1) = onnx::Reshape(%729, %742) # <ipython-input-6-cc76a8bf4ad5>:199:0\n  %yolo1 : Float(1:3360, 3:1120, 10:112, 16:7, 7:1) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%743) # <ipython-input-6-cc76a8bf4ad5>:219:0\n  %745 : Tensor = onnx::Constant[value={4}]()\n  %746 : Tensor = onnx::Constant[value={0}]()\n  %747 : Tensor = onnx::Constant[value={2}]()\n  %748 : Tensor = onnx::Constant[value={1}]()\n  %749 : Float(1:3360, 3:1120, 10:112, 16:7, 2:1) = onnx::Slice(%yolo1, %746, %747, %745, %748) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %750 : Float(1:960, 3:320, 10:32, 16:2, 2:1) = onnx::Sigmoid(%749) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %751 : Float(1:320, 1:320, 10:32, 16:2, 2:1) = onnx::Constant[value=<Tensor>]()\n  %752 : Float(1:960, 3:320, 10:32, 16:2, 2:1) = onnx::Add(%750, %751) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %753 : Tensor = onnx::Constant[value=  3  10  16   2 [ CPULongType{4} ]]()\n  %754 : Float(3:320, 10:32, 16:2, 2:1) = onnx::Reshape(%752, %753) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %755 : Tensor = onnx::Constant[value=  1   3  10  16   2 [ CPULongType{5} ]]()\n  %757 : Tensor = onnx::ConstantOfShape[value={1}](%1540)\n  %758 : Long() = onnx::Constant[value={-1}]()\n  %759 : LongTensor = onnx::Mul(%757, %758)\n  %760 : Long(5:1) = onnx::Constant[value=  1   3  10  16   2 [ CPULongType{5} ]]()\n  %761 : Tensor = onnx::Equal(%760, %759)\n  %762 : Tensor = onnx::Where(%761, %757, %755)\n  %763 : Float(1:960, 3:320, 10:32, 16:2, 2:1) = onnx::Expand(%754, %762) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %764 : Tensor = onnx::Shape(%yolo1)\n  %765 : Tensor = onnx::Constant[value={0}]()\n  %766 : Long() = onnx::Gather[axis=0](%764, %765)\n  %767 : Tensor = onnx::Cast[to=7](%766)\n  %768 : Tensor = onnx::Constant[value={0}]()\n  %769 : Tensor = onnx::Constant[value={1}]()\n  %770 : Tensor = onnx::Range(%768, %767, %769)\n  %771 : Tensor = onnx::Shape(%yolo1)\n  %772 : Tensor = onnx::Constant[value={1}]()\n  %773 : Long() = onnx::Gather[axis=0](%771, %772)\n  %774 : Tensor = onnx::Cast[to=7](%773)\n  %775 : Tensor = onnx::Constant[value={0}]()\n  %776 : Tensor = onnx::Constant[value={1}]()\n  %777 : Tensor = onnx::Range(%775, %774, %776)\n  %778 : Tensor = onnx::Shape(%yolo1)\n  %779 : Tensor = onnx::Constant[value={2}]()\n  %780 : Long() = onnx::Gather[axis=0](%778, %779)\n  %781 : Tensor = onnx::Cast[to=7](%780)\n  %782 : Tensor = onnx::Constant[value={0}]()\n  %783 : Tensor = onnx::Constant[value={1}]()\n  %784 : Tensor = onnx::Range(%782, %781, %783)\n  %785 : Tensor = onnx::Shape(%yolo1)\n  %786 : Tensor = onnx::Constant[value={3}]()\n  %787 : Long() = onnx::Gather[axis=0](%785, %786)\n  %788 : Tensor = onnx::Cast[to=7](%787)\n  %789 : Tensor = onnx::Constant[value={0}]()\n  %790 : Tensor = onnx::Constant[value={1}]()\n  %791 : Tensor = onnx::Range(%789, %788, %790)\n  %792 : Tensor = onnx::Shape(%yolo1)\n  %793 : Tensor = onnx::Constant[value={4}]()\n  %794 : Long() = onnx::Gather[axis=0](%792, %793)\n  %795 : Tensor = onnx::Cast[to=7](%794)\n  %796 : Tensor = onnx::Constant[value={0}]()\n  %797 : Tensor = onnx::Constant[value={1}]()\n  %798 : Tensor = onnx::Range(%796, %795, %797)\n  %799 : Tensor = onnx::Constant[value={0}]()\n  %800 : Tensor = onnx::Constant[value={0}]()\n  %801 : Tensor = onnx::Constant[value={2}]()\n  %802 : Tensor = onnx::Constant[value={1}]()\n  %803 : Tensor = onnx::Slice(%798, %800, %801, %799, %802)\n  %804 : Tensor = onnx::Constant[value=-1  1  1  1  1 [ CPULongType{5} ]]()\n  %805 : Tensor = onnx::Reshape(%770, %804)\n  %806 : Tensor = onnx::Constant[value=-1  1  1  1 [ CPULongType{4} ]]()\n  %807 : Tensor = onnx::Reshape(%777, %806)\n  %808 : Tensor = onnx::Constant[value=-1  1  1 [ CPULongType{3} ]]()\n  %809 : Tensor = onnx::Reshape(%784, %808)\n  %810 : Tensor = onnx::Constant[value=-1  1 [ CPULongType{2} ]]()\n  %811 : Tensor = onnx::Reshape(%791, %810)\n  %812 : Tensor = onnx::Constant[value={-1}]()\n  %813 : Tensor = onnx::Reshape(%803, %812)\n  %814 : Tensor = onnx::Add(%805, %807)\n  %815 : Tensor = onnx::Add(%814, %809)\n  %816 : Tensor = onnx::Add(%815, %811)\n  %817 : Tensor = onnx::Add(%816, %813)\n  %818 : Tensor = onnx::Shape(%817)\n  %819 : Tensor = onnx::Shape(%818)\n  %820 : Tensor = onnx::ConstantOfShape[value={1}](%819)\n  %821 : Long() = onnx::Constant[value={-1}]()\n  %822 : LongTensor = onnx::Mul(%820, %821)\n  %823 : Tensor = onnx::Equal(%818, %822)\n  %824 : Tensor = onnx::Where(%823, %820, %818)\n  %825 : Tensor = onnx::Expand(%805, %824)\n  %826 : Tensor = onnx::Unsqueeze[axes=[-1]](%825)\n  %827 : Tensor = onnx::Shape(%818)\n  %828 : Tensor = onnx::ConstantOfShape[value={1}](%827)\n  %829 : Long() = onnx::Constant[value={-1}]()\n  %830 : LongTensor = onnx::Mul(%828, %829)\n  %831 : Tensor = onnx::Equal(%818, %830)\n  %832 : Tensor = onnx::Where(%831, %828, %818)\n  %833 : Tensor = onnx::Expand(%807, %832)\n  %834 : Tensor = onnx::Unsqueeze[axes=[-1]](%833)\n  %835 : Tensor = onnx::Shape(%818)\n  %836 : Tensor = onnx::ConstantOfShape[value={1}](%835)\n  %837 : Long() = onnx::Constant[value={-1}]()\n  %838 : LongTensor = onnx::Mul(%836, %837)\n  %839 : Tensor = onnx::Equal(%818, %838)\n  %840 : Tensor = onnx::Where(%839, %836, %818)\n  %841 : Tensor = onnx::Expand(%809, %840)\n  %842 : Tensor = onnx::Unsqueeze[axes=[-1]](%841)\n  %843 : Tensor = onnx::Shape(%818)\n  %844 : Tensor = onnx::ConstantOfShape[value={1}](%843)\n  %845 : Long() = onnx::Constant[value={-1}]()\n  %846 : LongTensor = onnx::Mul(%844, %845)\n  %847 : Tensor = onnx::Equal(%818, %846)\n  %848 : Tensor = onnx::Where(%847, %844, %818)\n  %849 : Tensor = onnx::Expand(%811, %848)\n  %850 : Tensor = onnx::Unsqueeze[axes=[-1]](%849)\n  %851 : Tensor = onnx::Shape(%818)\n  %852 : Tensor = onnx::ConstantOfShape[value={1}](%851)\n  %853 : Long() = onnx::Constant[value={-1}]()\n  %854 : LongTensor = onnx::Mul(%852, %853)\n  %855 : Tensor = onnx::Equal(%818, %854)\n  %856 : Tensor = onnx::Where(%855, %852, %818)\n  %857 : Tensor = onnx::Expand(%813, %856)\n  %858 : Tensor = onnx::Unsqueeze[axes=[-1]](%857)\n  %859 : Tensor = onnx::Concat[axis=-1](%826, %834, %842, %850, %858)\n  %860 : Tensor = onnx::Shape(%yolo1)\n  %861 : Tensor = onnx::Constant[value={0}]()\n  %862 : Tensor = onnx::Constant[value={5}]()\n  %863 : Tensor = onnx::Constant[value={9223372036854775807}]()\n  %864 : Tensor = onnx::Slice(%860, %862, %863, %861)\n  %865 : Tensor = onnx::Concat[axis=0](%818, %864)\n  %866 : Tensor = onnx::Reshape(%763, %865)\n  %867 : Float(1:3360, 3:1120, 10:112, 16:7, 2:1) = onnx::ScatterND(%yolo1, %859, %866)\n  %868 : Tensor = onnx::Constant[value={4}]()\n  %869 : Tensor = onnx::Constant[value={2}]()\n  %870 : Tensor = onnx::Constant[value={4}]()\n  %871 : Tensor = onnx::Constant[value={1}]()\n  %872 : Float(1:3360, 3:1120, 10:112, 16:7, 2:1) = onnx::Slice(%867, %869, %870, %868, %871) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %873 : Float(1:960, 3:320, 10:32, 16:2, 2:1) = onnx::Exp(%872) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %874 : Float(1:6, 3:2, 1:2, 1:2, 2:1) = onnx::Constant[value=(1,1,1,.,.) =    3.5938  2.2812  (1,2,1,.,.) =    3.7188  6.2188  (1,3,1,.,.) =    7.5625  7.4375 [ CPUFloatType{1,3,1,1,2} ]]()\n  %875 : Float(1:960, 3:320, 10:32, 16:2, 2:1) = onnx::Mul(%873, %874) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %876 : Tensor = onnx::Constant[value=  3  10  16   2 [ CPULongType{4} ]]()\n  %877 : Float(3:320, 10:32, 16:2, 2:1) = onnx::Reshape(%875, %876) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %878 : Tensor = onnx::Constant[value=  1   3  10  16   2 [ CPULongType{5} ]]()\n  %880 : Tensor = onnx::ConstantOfShape[value={1}](%1541)\n  %881 : Long() = onnx::Constant[value={-1}]()\n  %882 : LongTensor = onnx::Mul(%880, %881)\n  %883 : Long(5:1) = onnx::Constant[value=  1   3  10  16   2 [ CPULongType{5} ]]()\n  %884 : Tensor = onnx::Equal(%883, %882)\n  %885 : Tensor = onnx::Where(%884, %880, %878)\n  %886 : Float(1:960, 3:320, 10:32, 16:2, 2:1) = onnx::Expand(%877, %885) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %887 : Tensor = onnx::Shape(%867)\n  %888 : Tensor = onnx::Constant[value={0}]()\n  %889 : Long() = onnx::Gather[axis=0](%887, %888)\n  %890 : Tensor = onnx::Cast[to=7](%889)\n  %891 : Tensor = onnx::Constant[value={0}]()\n  %892 : Tensor = onnx::Constant[value={1}]()\n  %893 : Tensor = onnx::Range(%891, %890, %892)\n  %894 : Tensor = onnx::Shape(%867)\n  %895 : Tensor = onnx::Constant[value={1}]()\n  %896 : Long() = onnx::Gather[axis=0](%894, %895)\n  %897 : Tensor = onnx::Cast[to=7](%896)\n  %898 : Tensor = onnx::Constant[value={0}]()\n  %899 : Tensor = onnx::Constant[value={1}]()\n  %900 : Tensor = onnx::Range(%898, %897, %899)\n  %901 : Tensor = onnx::Shape(%867)\n  %902 : Tensor = onnx::Constant[value={2}]()\n  %903 : Long() = onnx::Gather[axis=0](%901, %902)\n  %904 : Tensor = onnx::Cast[to=7](%903)\n  %905 : Tensor = onnx::Constant[value={0}]()\n  %906 : Tensor = onnx::Constant[value={1}]()\n  %907 : Tensor = onnx::Range(%905, %904, %906)\n  %908 : Tensor = onnx::Shape(%867)\n  %909 : Tensor = onnx::Constant[value={3}]()\n  %910 : Long() = onnx::Gather[axis=0](%908, %909)\n  %911 : Tensor = onnx::Cast[to=7](%910)\n  %912 : Tensor = onnx::Constant[value={0}]()\n  %913 : Tensor = onnx::Constant[value={1}]()\n  %914 : Tensor = onnx::Range(%912, %911, %913)\n  %915 : Tensor = onnx::Shape(%867)\n  %916 : Tensor = onnx::Constant[value={4}]()\n  %917 : Long() = onnx::Gather[axis=0](%915, %916)\n  %918 : Tensor = onnx::Cast[to=7](%917)\n  %919 : Tensor = onnx::Constant[value={0}]()\n  %920 : Tensor = onnx::Constant[value={1}]()\n  %921 : Tensor = onnx::Range(%919, %918, %920)\n  %922 : Tensor = onnx::Constant[value={0}]()\n  %923 : Tensor = onnx::Constant[value={2}]()\n  %924 : Tensor = onnx::Constant[value={4}]()\n  %925 : Tensor = onnx::Constant[value={1}]()\n  %926 : Tensor = onnx::Slice(%921, %923, %924, %922, %925)\n  %927 : Tensor = onnx::Constant[value=-1  1  1  1  1 [ CPULongType{5} ]]()\n  %928 : Tensor = onnx::Reshape(%893, %927)\n  %929 : Tensor = onnx::Constant[value=-1  1  1  1 [ CPULongType{4} ]]()\n  %930 : Tensor = onnx::Reshape(%900, %929)\n  %931 : Tensor = onnx::Constant[value=-1  1  1 [ CPULongType{3} ]]()\n  %932 : Tensor = onnx::Reshape(%907, %931)\n  %933 : Tensor = onnx::Constant[value=-1  1 [ CPULongType{2} ]]()\n  %934 : Tensor = onnx::Reshape(%914, %933)\n  %935 : Tensor = onnx::Constant[value={-1}]()\n  %936 : Tensor = onnx::Reshape(%926, %935)\n  %937 : Tensor = onnx::Add(%928, %930)\n  %938 : Tensor = onnx::Add(%937, %932)\n  %939 : Tensor = onnx::Add(%938, %934)\n  %940 : Tensor = onnx::Add(%939, %936)\n  %941 : Tensor = onnx::Shape(%940)\n  %942 : Tensor = onnx::Shape(%941)\n  %943 : Tensor = onnx::ConstantOfShape[value={1}](%942)\n  %944 : Long() = onnx::Constant[value={-1}]()\n  %945 : LongTensor = onnx::Mul(%943, %944)\n  %946 : Tensor = onnx::Equal(%941, %945)\n  %947 : Tensor = onnx::Where(%946, %943, %941)\n  %948 : Tensor = onnx::Expand(%928, %947)\n  %949 : Tensor = onnx::Unsqueeze[axes=[-1]](%948)\n  %950 : Tensor = onnx::Shape(%941)\n  %951 : Tensor = onnx::ConstantOfShape[value={1}](%950)\n  %952 : Long() = onnx::Constant[value={-1}]()\n  %953 : LongTensor = onnx::Mul(%951, %952)\n  %954 : Tensor = onnx::Equal(%941, %953)\n  %955 : Tensor = onnx::Where(%954, %951, %941)\n  %956 : Tensor = onnx::Expand(%930, %955)\n  %957 : Tensor = onnx::Unsqueeze[axes=[-1]](%956)\n  %958 : Tensor = onnx::Shape(%941)\n  %959 : Tensor = onnx::ConstantOfShape[value={1}](%958)\n  %960 : Long() = onnx::Constant[value={-1}]()\n  %961 : LongTensor = onnx::Mul(%959, %960)\n  %962 : Tensor = onnx::Equal(%941, %961)\n  %963 : Tensor = onnx::Where(%962, %959, %941)\n  %964 : Tensor = onnx::Expand(%932, %963)\n  %965 : Tensor = onnx::Unsqueeze[axes=[-1]](%964)\n  %966 : Tensor = onnx::Shape(%941)\n  %967 : Tensor = onnx::ConstantOfShape[value={1}](%966)\n  %968 : Long() = onnx::Constant[value={-1}]()\n  %969 : LongTensor = onnx::Mul(%967, %968)\n  %970 : Tensor = onnx::Equal(%941, %969)\n  %971 : Tensor = onnx::Where(%970, %967, %941)\n  %972 : Tensor = onnx::Expand(%934, %971)\n  %973 : Tensor = onnx::Unsqueeze[axes=[-1]](%972)\n  %974 : Tensor = onnx::Shape(%941)\n  %975 : Tensor = onnx::ConstantOfShape[value={1}](%974)\n  %976 : Long() = onnx::Constant[value={-1}]()\n  %977 : LongTensor = onnx::Mul(%975, %976)\n  %978 : Tensor = onnx::Equal(%941, %977)\n  %979 : Tensor = onnx::Where(%978, %975, %941)\n  %980 : Tensor = onnx::Expand(%936, %979)\n  %981 : Tensor = onnx::Unsqueeze[axes=[-1]](%980)\n  %982 : Tensor = onnx::Concat[axis=-1](%949, %957, %965, %973, %981)\n  %983 : Tensor = onnx::Shape(%867)\n  %984 : Tensor = onnx::Constant[value={0}]()\n  %985 : Tensor = onnx::Constant[value={5}]()\n  %986 : Tensor = onnx::Constant[value={9223372036854775807}]()\n  %987 : Tensor = onnx::Slice(%983, %985, %986, %984)\n  %988 : Tensor = onnx::Concat[axis=0](%941, %987)\n  %989 : Tensor = onnx::Reshape(%886, %988)\n  %990 : Float(1:3360, 3:1120, 10:112, 16:7, 2:1) = onnx::ScatterND(%867, %982, %989)\n  %991 : Tensor = onnx::Constant[value={4}]()\n  %992 : Tensor = onnx::Constant[value={0}]()\n  %993 : Tensor = onnx::Constant[value={4}]()\n  %994 : Tensor = onnx::Constant[value={1}]()\n  %995 : Float(1:3360, 3:1120, 10:112, 16:7, 4:1) = onnx::Slice(%990, %992, %993, %991, %994) # <ipython-input-6-cc76a8bf4ad5>:222:0\n  %996 : Float() = onnx::Constant[value={32}]()\n  %997 : Float(1:3360, 3:1120, 10:112, 16:7, 4:1) = onnx::Mul(%995, %996)\n  %998 : Tensor = onnx::Constant[value=  3  10  16   4 [ CPULongType{4} ]]()\n  %999 : Float(3:1120, 10:112, 16:7, 4:1) = onnx::Reshape(%997, %998) # <ipython-input-6-cc76a8bf4ad5>:222:0\n  %1000 : Tensor = onnx::Constant[value=  1   3  10  16   4 [ CPULongType{5} ]]()\n  %1002 : Tensor = onnx::ConstantOfShape[value={1}](%1542)\n  %1003 : Long() = onnx::Constant[value={-1}]()\n  %1004 : LongTensor = onnx::Mul(%1002, %1003)\n  %1005 : Long(5:1) = onnx::Constant[value=  1   3  10  16   4 [ CPULongType{5} ]]()\n  %1006 : Tensor = onnx::Equal(%1005, %1004)\n  %1007 : Tensor = onnx::Where(%1006, %1002, %1000)\n  %1008 : Float(1:3360, 3:1120, 10:112, 16:7, 4:1) = onnx::Expand(%999, %1007) # <ipython-input-6-cc76a8bf4ad5>:222:0\n  %1009 : Tensor = onnx::Shape(%990)\n  %1010 : Tensor = onnx::Constant[value={0}]()\n  %1011 : Long() = onnx::Gather[axis=0](%1009, %1010)\n  %1012 : Tensor = onnx::Cast[to=7](%1011)\n  %1013 : Tensor = onnx::Constant[value={0}]()\n  %1014 : Tensor = onnx::Constant[value={1}]()\n  %1015 : Tensor = onnx::Range(%1013, %1012, %1014)\n  %1016 : Tensor = onnx::Shape(%990)\n  %1017 : Tensor = onnx::Constant[value={1}]()\n  %1018 : Long() = onnx::Gather[axis=0](%1016, %1017)\n  %1019 : Tensor = onnx::Cast[to=7](%1018)\n  %1020 : Tensor = onnx::Constant[value={0}]()\n  %1021 : Tensor = onnx::Constant[value={1}]()\n  %1022 : Tensor = onnx::Range(%1020, %1019, %1021)\n  %1023 : Tensor = onnx::Shape(%990)\n  %1024 : Tensor = onnx::Constant[value={2}]()\n  %1025 : Long() = onnx::Gather[axis=0](%1023, %1024)\n  %1026 : Tensor = onnx::Cast[to=7](%1025)\n  %1027 : Tensor = onnx::Constant[value={0}]()\n  %1028 : Tensor = onnx::Constant[value={1}]()\n  %1029 : Tensor = onnx::Range(%1027, %1026, %1028)\n  %1030 : Tensor = onnx::Shape(%990)\n  %1031 : Tensor = onnx::Constant[value={3}]()\n  %1032 : Long() = onnx::Gather[axis=0](%1030, %1031)\n  %1033 : Tensor = onnx::Cast[to=7](%1032)\n  %1034 : Tensor = onnx::Constant[value={0}]()\n  %1035 : Tensor = onnx::Constant[value={1}]()\n  %1036 : Tensor = onnx::Range(%1034, %1033, %1035)\n  %1037 : Tensor = onnx::Shape(%990)\n  %1038 : Tensor = onnx::Constant[value={4}]()\n  %1039 : Long() = onnx::Gather[axis=0](%1037, %1038)\n  %1040 : Tensor = onnx::Cast[to=7](%1039)\n  %1041 : Tensor = onnx::Constant[value={0}]()\n  %1042 : Tensor = onnx::Constant[value={1}]()\n  %1043 : Tensor = onnx::Range(%1041, %1040, %1042)\n  %1044 : Tensor = onnx::Constant[value={0}]()\n  %1045 : Tensor = onnx::Constant[value={0}]()\n  %1046 : Tensor = onnx::Constant[value={4}]()\n  %1047 : Tensor = onnx::Constant[value={1}]()\n  %1048 : Tensor = onnx::Slice(%1043, %1045, %1046, %1044, %1047)\n  %1049 : Tensor = onnx::Constant[value=-1  1  1  1  1 [ CPULongType{5} ]]()\n  %1050 : Tensor = onnx::Reshape(%1015, %1049)\n  %1051 : Tensor = onnx::Constant[value=-1  1  1  1 [ CPULongType{4} ]]()\n  %1052 : Tensor = onnx::Reshape(%1022, %1051)\n  %1053 : Tensor = onnx::Constant[value=-1  1  1 [ CPULongType{3} ]]()\n  %1054 : Tensor = onnx::Reshape(%1029, %1053)\n  %1055 : Tensor = onnx::Constant[value=-1  1 [ CPULongType{2} ]]()\n  %1056 : Tensor = onnx::Reshape(%1036, %1055)\n  %1057 : Tensor = onnx::Constant[value={-1}]()\n  %1058 : Tensor = onnx::Reshape(%1048, %1057)\n  %1059 : Tensor = onnx::Add(%1050, %1052)\n  %1060 : Tensor = onnx::Add(%1059, %1054)\n  %1061 : Tensor = onnx::Add(%1060, %1056)\n  %1062 : Tensor = onnx::Add(%1061, %1058)\n  %1063 : Tensor = onnx::Shape(%1062)\n  %1064 : Tensor = onnx::Shape(%1063)\n  %1065 : Tensor = onnx::ConstantOfShape[value={1}](%1064)\n  %1066 : Long() = onnx::Constant[value={-1}]()\n  %1067 : LongTensor = onnx::Mul(%1065, %1066)\n  %1068 : Tensor = onnx::Equal(%1063, %1067)\n  %1069 : Tensor = onnx::Where(%1068, %1065, %1063)\n  %1070 : Tensor = onnx::Expand(%1050, %1069)\n  %1071 : Tensor = onnx::Unsqueeze[axes=[-1]](%1070)\n  %1072 : Tensor = onnx::Shape(%1063)\n  %1073 : Tensor = onnx::ConstantOfShape[value={1}](%1072)\n  %1074 : Long() = onnx::Constant[value={-1}]()\n  %1075 : LongTensor = onnx::Mul(%1073, %1074)\n  %1076 : Tensor = onnx::Equal(%1063, %1075)\n  %1077 : Tensor = onnx::Where(%1076, %1073, %1063)\n  %1078 : Tensor = onnx::Expand(%1052, %1077)\n  %1079 : Tensor = onnx::Unsqueeze[axes=[-1]](%1078)\n  %1080 : Tensor = onnx::Shape(%1063)\n  %1081 : Tensor = onnx::ConstantOfShape[value={1}](%1080)\n  %1082 : Long() = onnx::Constant[value={-1}]()\n  %1083 : LongTensor = onnx::Mul(%1081, %1082)\n  %1084 : Tensor = onnx::Equal(%1063, %1083)\n  %1085 : Tensor = onnx::Where(%1084, %1081, %1063)\n  %1086 : Tensor = onnx::Expand(%1054, %1085)\n  %1087 : Tensor = onnx::Unsqueeze[axes=[-1]](%1086)\n  %1088 : Tensor = onnx::Shape(%1063)\n  %1089 : Tensor = onnx::ConstantOfShape[value={1}](%1088)\n  %1090 : Long() = onnx::Constant[value={-1}]()\n  %1091 : LongTensor = onnx::Mul(%1089, %1090)\n  %1092 : Tensor = onnx::Equal(%1063, %1091)\n  %1093 : Tensor = onnx::Where(%1092, %1089, %1063)\n  %1094 : Tensor = onnx::Expand(%1056, %1093)\n  %1095 : Tensor = onnx::Unsqueeze[axes=[-1]](%1094)\n  %1096 : Tensor = onnx::Shape(%1063)\n  %1097 : Tensor = onnx::ConstantOfShape[value={1}](%1096)\n  %1098 : Long() = onnx::Constant[value={-1}]()\n  %1099 : LongTensor = onnx::Mul(%1097, %1098)\n  %1100 : Tensor = onnx::Equal(%1063, %1099)\n  %1101 : Tensor = onnx::Where(%1100, %1097, %1063)\n  %1102 : Tensor = onnx::Expand(%1058, %1101)\n  %1103 : Tensor = onnx::Unsqueeze[axes=[-1]](%1102)\n  %1104 : Tensor = onnx::Concat[axis=-1](%1071, %1079, %1087, %1095, %1103)\n  %1105 : Tensor = onnx::Shape(%990)\n  %1106 : Tensor = onnx::Constant[value={0}]()\n  %1107 : Tensor = onnx::Constant[value={5}]()\n  %1108 : Tensor = onnx::Constant[value={9223372036854775807}]()\n  %1109 : Tensor = onnx::Slice(%1105, %1107, %1108, %1106)\n  %1110 : Tensor = onnx::Concat[axis=0](%1063, %1109)\n  %1111 : Tensor = onnx::Reshape(%1008, %1110)\n  %1112 : Float(1:3360, 3:1120, 10:112, 16:7, 4:1) = onnx::ScatterND(%990, %1104, %1111)\n  %1115 : Tensor = onnx::Unsqueeze[axes=[0]](%732)\n  %1118 : Tensor = onnx::Concat[axis=0](%1115, %1543, %1544)\n  %1119 : Float(1:3360, 480:7, 7:1) = onnx::Reshape(%1112, %1118) # <ipython-input-6-cc76a8bf4ad5>:224:0\n  %1128 : Tensor = onnx::Constant[value=[ CPUFloatType{0} ]]()\n  %1129 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Resize[coordinate_transformation_mode=\"asymmetric\", cubic_coeff_a=-0.75, mode=\"nearest\", nearest_mode=\"floor\"](%718, %1128, %1549) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:3143:0\n  %1130 : Float(1:148480, 232:640, 20:32, 32:1) = onnx::Concat[axis=1](%1129, %665) # /Users/waittim/Google 云端硬盘/Mask-detection/mask-detector/modeling/utils/layers.py:35:0\n  %1131 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%1130, %module_list.119.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %1132 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%1131, %module_list.119.BatchNorm2d.weight, %module_list.119.BatchNorm2d.bias, %module_list.119.BatchNorm2d.running_mean, %module_list.119.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %1133 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%1132) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %1134 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[5, 5], pads=[2, 2, 2, 2], strides=[1, 1]](%1133, %module_list.120.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %1135 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%1134, %module_list.120.BatchNorm2d.weight, %module_list.120.BatchNorm2d.bias, %module_list.120.BatchNorm2d.running_mean, %module_list.120.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %1136 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%1135) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %1137 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%1136, %module_list.121.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %1138 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%1137, %module_list.121.BatchNorm2d.weight, %module_list.121.BatchNorm2d.bias, %module_list.121.BatchNorm2d.running_mean, %module_list.121.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %1139 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=96, kernel_shape=[5, 5], pads=[2, 2, 2, 2], strides=[1, 1]](%1138, %module_list.122.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %1140 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%1139, %module_list.122.BatchNorm2d.weight, %module_list.122.BatchNorm2d.bias, %module_list.122.BatchNorm2d.running_mean, %module_list.122.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %1141 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::LeakyRelu[alpha=0.10000000000000001](%1140) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:1293:0\n  %1142 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%1141, %module_list.123.Conv2d.weight) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %1143 : Float(1:61440, 96:640, 20:32, 32:1) = onnx::BatchNormalization[epsilon=0.0001, momentum=0.96999999999999997](%1142, %module_list.123.BatchNorm2d.weight, %module_list.123.BatchNorm2d.bias, %module_list.123.BatchNorm2d.running_mean, %module_list.123.BatchNorm2d.running_var) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/functional.py:2016:0\n  %1144 : Float(1:13440, 21:640, 20:32, 32:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%1143, %module_list.124.Conv2d.weight, %module_list.124.Conv2d.bias) # /Users/waittim/anaconda3/lib/python3.7/site-packages/torch/nn/modules/conv.py:416:0\n  %1145 : Tensor = onnx::Shape(%1144)\n  %1146 : Tensor = onnx::Constant[value={0}]()\n  %1147 : Long() = onnx::Gather[axis=0](%1145, %1146) # <ipython-input-6-cc76a8bf4ad5>:194:0\n  %1152 : Tensor = onnx::Unsqueeze[axes=[0]](%1147)\n  %1157 : Tensor = onnx::Concat[axis=0](%1152, %1550, %1551, %1552, %1553)\n  %1158 : Float(1:13440, 3:4480, 7:640, 20:32, 32:1) = onnx::Reshape(%1144, %1157) # <ipython-input-6-cc76a8bf4ad5>:199:0\n  %yolo2 : Float(1:13440, 3:4480, 20:224, 32:7, 7:1) = onnx::Transpose[perm=[0, 1, 3, 4, 2]](%1158) # <ipython-input-6-cc76a8bf4ad5>:219:0\n  %1160 : Tensor = onnx::Constant[value={4}]()\n  %1161 : Tensor = onnx::Constant[value={0}]()\n  %1162 : Tensor = onnx::Constant[value={2}]()\n  %1163 : Tensor = onnx::Constant[value={1}]()\n  %1164 : Float(1:13440, 3:4480, 20:224, 32:7, 2:1) = onnx::Slice(%yolo2, %1161, %1162, %1160, %1163) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %1165 : Float(1:3840, 3:1280, 20:64, 32:2, 2:1) = onnx::Sigmoid(%1164) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %1166 : Float(1:1280, 1:1280, 20:64, 32:2, 2:1) = onnx::Constant[value=<Tensor>]()\n  %1167 : Float(1:3840, 3:1280, 20:64, 32:2, 2:1) = onnx::Add(%1165, %1166) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %1168 : Tensor = onnx::Constant[value=  3  20  32   2 [ CPULongType{4} ]]()\n  %1169 : Float(3:1280, 20:64, 32:2, 2:1) = onnx::Reshape(%1167, %1168) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %1170 : Tensor = onnx::Constant[value=  1   3  20  32   2 [ CPULongType{5} ]]()\n  %1172 : Tensor = onnx::ConstantOfShape[value={1}](%1554)\n  %1173 : Long() = onnx::Constant[value={-1}]()\n  %1174 : LongTensor = onnx::Mul(%1172, %1173)\n  %1175 : Long(5:1) = onnx::Constant[value=  1   3  20  32   2 [ CPULongType{5} ]]()\n  %1176 : Tensor = onnx::Equal(%1175, %1174)\n  %1177 : Tensor = onnx::Where(%1176, %1172, %1170)\n  %1178 : Float(1:3840, 3:1280, 20:64, 32:2, 2:1) = onnx::Expand(%1169, %1177) # <ipython-input-6-cc76a8bf4ad5>:220:0\n  %1179 : Tensor = onnx::Shape(%yolo2)\n  %1180 : Tensor = onnx::Constant[value={0}]()\n  %1181 : Long() = onnx::Gather[axis=0](%1179, %1180)\n  %1182 : Tensor = onnx::Cast[to=7](%1181)\n  %1183 : Tensor = onnx::Constant[value={0}]()\n  %1184 : Tensor = onnx::Constant[value={1}]()\n  %1185 : Tensor = onnx::Range(%1183, %1182, %1184)\n  %1186 : Tensor = onnx::Shape(%yolo2)\n  %1187 : Tensor = onnx::Constant[value={1}]()\n  %1188 : Long() = onnx::Gather[axis=0](%1186, %1187)\n  %1189 : Tensor = onnx::Cast[to=7](%1188)\n  %1190 : Tensor = onnx::Constant[value={0}]()\n  %1191 : Tensor = onnx::Constant[value={1}]()\n  %1192 : Tensor = onnx::Range(%1190, %1189, %1191)\n  %1193 : Tensor = onnx::Shape(%yolo2)\n  %1194 : Tensor = onnx::Constant[value={2}]()\n  %1195 : Long() = onnx::Gather[axis=0](%1193, %1194)\n  %1196 : Tensor = onnx::Cast[to=7](%1195)\n  %1197 : Tensor = onnx::Constant[value={0}]()\n  %1198 : Tensor = onnx::Constant[value={1}]()\n  %1199 : Tensor = onnx::Range(%1197, %1196, %1198)\n  %1200 : Tensor = onnx::Shape(%yolo2)\n  %1201 : Tensor = onnx::Constant[value={3}]()\n  %1202 : Long() = onnx::Gather[axis=0](%1200, %1201)\n  %1203 : Tensor = onnx::Cast[to=7](%1202)\n  %1204 : Tensor = onnx::Constant[value={0}]()\n  %1205 : Tensor = onnx::Constant[value={1}]()\n  %1206 : Tensor = onnx::Range(%1204, %1203, %1205)\n  %1207 : Tensor = onnx::Shape(%yolo2)\n  %1208 : Tensor = onnx::Constant[value={4}]()\n  %1209 : Long() = onnx::Gather[axis=0](%1207, %1208)\n  %1210 : Tensor = onnx::Cast[to=7](%1209)\n  %1211 : Tensor = onnx::Constant[value={0}]()\n  %1212 : Tensor = onnx::Constant[value={1}]()\n  %1213 : Tensor = onnx::Range(%1211, %1210, %1212)\n  %1214 : Tensor = onnx::Constant[value={0}]()\n  %1215 : Tensor = onnx::Constant[value={0}]()\n  %1216 : Tensor = onnx::Constant[value={2}]()\n  %1217 : Tensor = onnx::Constant[value={1}]()\n  %1218 : Tensor = onnx::Slice(%1213, %1215, %1216, %1214, %1217)\n  %1219 : Tensor = onnx::Constant[value=-1  1  1  1  1 [ CPULongType{5} ]]()\n  %1220 : Tensor = onnx::Reshape(%1185, %1219)\n  %1221 : Tensor = onnx::Constant[value=-1  1  1  1 [ CPULongType{4} ]]()\n  %1222 : Tensor = onnx::Reshape(%1192, %1221)\n  %1223 : Tensor = onnx::Constant[value=-1  1  1 [ CPULongType{3} ]]()\n  %1224 : Tensor = onnx::Reshape(%1199, %1223)\n  %1225 : Tensor = onnx::Constant[value=-1  1 [ CPULongType{2} ]]()\n  %1226 : Tensor = onnx::Reshape(%1206, %1225)\n  %1227 : Tensor = onnx::Constant[value={-1}]()\n  %1228 : Tensor = onnx::Reshape(%1218, %1227)\n  %1229 : Tensor = onnx::Add(%1220, %1222)\n  %1230 : Tensor = onnx::Add(%1229, %1224)\n  %1231 : Tensor = onnx::Add(%1230, %1226)\n  %1232 : Tensor = onnx::Add(%1231, %1228)\n  %1233 : Tensor = onnx::Shape(%1232)\n  %1234 : Tensor = onnx::Shape(%1233)\n  %1235 : Tensor = onnx::ConstantOfShape[value={1}](%1234)\n  %1236 : Long() = onnx::Constant[value={-1}]()\n  %1237 : LongTensor = onnx::Mul(%1235, %1236)\n  %1238 : Tensor = onnx::Equal(%1233, %1237)\n  %1239 : Tensor = onnx::Where(%1238, %1235, %1233)\n  %1240 : Tensor = onnx::Expand(%1220, %1239)\n  %1241 : Tensor = onnx::Unsqueeze[axes=[-1]](%1240)\n  %1242 : Tensor = onnx::Shape(%1233)\n  %1243 : Tensor = onnx::ConstantOfShape[value={1}](%1242)\n  %1244 : Long() = onnx::Constant[value={-1}]()\n  %1245 : LongTensor = onnx::Mul(%1243, %1244)\n  %1246 : Tensor = onnx::Equal(%1233, %1245)\n  %1247 : Tensor = onnx::Where(%1246, %1243, %1233)\n  %1248 : Tensor = onnx::Expand(%1222, %1247)\n  %1249 : Tensor = onnx::Unsqueeze[axes=[-1]](%1248)\n  %1250 : Tensor = onnx::Shape(%1233)\n  %1251 : Tensor = onnx::ConstantOfShape[value={1}](%1250)\n  %1252 : Long() = onnx::Constant[value={-1}]()\n  %1253 : LongTensor = onnx::Mul(%1251, %1252)\n  %1254 : Tensor = onnx::Equal(%1233, %1253)\n  %1255 : Tensor = onnx::Where(%1254, %1251, %1233)\n  %1256 : Tensor = onnx::Expand(%1224, %1255)\n  %1257 : Tensor = onnx::Unsqueeze[axes=[-1]](%1256)\n  %1258 : Tensor = onnx::Shape(%1233)\n  %1259 : Tensor = onnx::ConstantOfShape[value={1}](%1258)\n  %1260 : Long() = onnx::Constant[value={-1}]()\n  %1261 : LongTensor = onnx::Mul(%1259, %1260)\n  %1262 : Tensor = onnx::Equal(%1233, %1261)\n  %1263 : Tensor = onnx::Where(%1262, %1259, %1233)\n  %1264 : Tensor = onnx::Expand(%1226, %1263)\n  %1265 : Tensor = onnx::Unsqueeze[axes=[-1]](%1264)\n  %1266 : Tensor = onnx::Shape(%1233)\n  %1267 : Tensor = onnx::ConstantOfShape[value={1}](%1266)\n  %1268 : Long() = onnx::Constant[value={-1}]()\n  %1269 : LongTensor = onnx::Mul(%1267, %1268)\n  %1270 : Tensor = onnx::Equal(%1233, %1269)\n  %1271 : Tensor = onnx::Where(%1270, %1267, %1233)\n  %1272 : Tensor = onnx::Expand(%1228, %1271)\n  %1273 : Tensor = onnx::Unsqueeze[axes=[-1]](%1272)\n  %1274 : Tensor = onnx::Concat[axis=-1](%1241, %1249, %1257, %1265, %1273)\n  %1275 : Tensor = onnx::Shape(%yolo2)\n  %1276 : Tensor = onnx::Constant[value={0}]()\n  %1277 : Tensor = onnx::Constant[value={5}]()\n  %1278 : Tensor = onnx::Constant[value={9223372036854775807}]()\n  %1279 : Tensor = onnx::Slice(%1275, %1277, %1278, %1276)\n  %1280 : Tensor = onnx::Concat[axis=0](%1233, %1279)\n  %1281 : Tensor = onnx::Reshape(%1178, %1280)\n  %1282 : Float(1:13440, 3:4480, 20:224, 32:7, 2:1) = onnx::ScatterND(%yolo2, %1274, %1281)\n  %1283 : Tensor = onnx::Constant[value={4}]()\n  %1284 : Tensor = onnx::Constant[value={2}]()\n  %1285 : Tensor = onnx::Constant[value={4}]()\n  %1286 : Tensor = onnx::Constant[value={1}]()\n  %1287 : Float(1:13440, 3:4480, 20:224, 32:7, 2:1) = onnx::Slice(%1282, %1284, %1285, %1283, %1286) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %1288 : Float(1:3840, 3:1280, 20:64, 32:2, 2:1) = onnx::Exp(%1287) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %1289 : Float(1:6, 3:2, 1:2, 1:2, 2:1) = onnx::Constant[value=(1,1,1,.,.) =    0.7500  1.1250  (1,2,1,.,.) =    2.3125  3.0625  (1,3,1,.,.) =    3.2500  8.2500 [ CPUFloatType{1,3,1,1,2} ]]()\n  %1290 : Float(1:3840, 3:1280, 20:64, 32:2, 2:1) = onnx::Mul(%1288, %1289) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %1291 : Tensor = onnx::Constant[value=  3  20  32   2 [ CPULongType{4} ]]()\n  %1292 : Float(3:1280, 20:64, 32:2, 2:1) = onnx::Reshape(%1290, %1291) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %1293 : Tensor = onnx::Constant[value=  1   3  20  32   2 [ CPULongType{5} ]]()\n  %1295 : Tensor = onnx::ConstantOfShape[value={1}](%1555)\n  %1296 : Long() = onnx::Constant[value={-1}]()\n  %1297 : LongTensor = onnx::Mul(%1295, %1296)\n  %1298 : Long(5:1) = onnx::Constant[value=  1   3  20  32   2 [ CPULongType{5} ]]()\n  %1299 : Tensor = onnx::Equal(%1298, %1297)\n  %1300 : Tensor = onnx::Where(%1299, %1295, %1293)\n  %1301 : Float(1:3840, 3:1280, 20:64, 32:2, 2:1) = onnx::Expand(%1292, %1300) # <ipython-input-6-cc76a8bf4ad5>:221:0\n  %1302 : Tensor = onnx::Shape(%1282)\n  %1303 : Tensor = onnx::Constant[value={0}]()\n  %1304 : Long() = onnx::Gather[axis=0](%1302, %1303)\n  %1305 : Tensor = onnx::Cast[to=7](%1304)\n  %1306 : Tensor = onnx::Constant[value={0}]()\n  %1307 : Tensor = onnx::Constant[value={1}]()\n  %1308 : Tensor = onnx::Range(%1306, %1305, %1307)\n  %1309 : Tensor = onnx::Shape(%1282)\n  %1310 : Tensor = onnx::Constant[value={1}]()\n  %1311 : Long() = onnx::Gather[axis=0](%1309, %1310)\n  %1312 : Tensor = onnx::Cast[to=7](%1311)\n  %1313 : Tensor = onnx::Constant[value={0}]()\n  %1314 : Tensor = onnx::Constant[value={1}]()\n  %1315 : Tensor = onnx::Range(%1313, %1312, %1314)\n  %1316 : Tensor = onnx::Shape(%1282)\n  %1317 : Tensor = onnx::Constant[value={2}]()\n  %1318 : Long() = onnx::Gather[axis=0](%1316, %1317)\n  %1319 : Tensor = onnx::Cast[to=7](%1318)\n  %1320 : Tensor = onnx::Constant[value={0}]()\n  %1321 : Tensor = onnx::Constant[value={1}]()\n  %1322 : Tensor = onnx::Range(%1320, %1319, %1321)\n  %1323 : Tensor = onnx::Shape(%1282)\n  %1324 : Tensor = onnx::Constant[value={3}]()\n  %1325 : Long() = onnx::Gather[axis=0](%1323, %1324)\n  %1326 : Tensor = onnx::Cast[to=7](%1325)\n  %1327 : Tensor = onnx::Constant[value={0}]()\n  %1328 : Tensor = onnx::Constant[value={1}]()\n  %1329 : Tensor = onnx::Range(%1327, %1326, %1328)\n  %1330 : Tensor = onnx::Shape(%1282)\n  %1331 : Tensor = onnx::Constant[value={4}]()\n  %1332 : Long() = onnx::Gather[axis=0](%1330, %1331)\n  %1333 : Tensor = onnx::Cast[to=7](%1332)\n  %1334 : Tensor = onnx::Constant[value={0}]()\n  %1335 : Tensor = onnx::Constant[value={1}]()\n  %1336 : Tensor = onnx::Range(%1334, %1333, %1335)\n  %1337 : Tensor = onnx::Constant[value={0}]()\n  %1338 : Tensor = onnx::Constant[value={2}]()\n  %1339 : Tensor = onnx::Constant[value={4}]()\n  %1340 : Tensor = onnx::Constant[value={1}]()\n  %1341 : Tensor = onnx::Slice(%1336, %1338, %1339, %1337, %1340)\n  %1342 : Tensor = onnx::Constant[value=-1  1  1  1  1 [ CPULongType{5} ]]()\n  %1343 : Tensor = onnx::Reshape(%1308, %1342)\n  %1344 : Tensor = onnx::Constant[value=-1  1  1  1 [ CPULongType{4} ]]()\n  %1345 : Tensor = onnx::Reshape(%1315, %1344)\n  %1346 : Tensor = onnx::Constant[value=-1  1  1 [ CPULongType{3} ]]()\n  %1347 : Tensor = onnx::Reshape(%1322, %1346)\n  %1348 : Tensor = onnx::Constant[value=-1  1 [ CPULongType{2} ]]()\n  %1349 : Tensor = onnx::Reshape(%1329, %1348)\n  %1350 : Tensor = onnx::Constant[value={-1}]()\n  %1351 : Tensor = onnx::Reshape(%1341, %1350)\n  %1352 : Tensor = onnx::Add(%1343, %1345)\n  %1353 : Tensor = onnx::Add(%1352, %1347)\n  %1354 : Tensor = onnx::Add(%1353, %1349)\n  %1355 : Tensor = onnx::Add(%1354, %1351)\n  %1356 : Tensor = onnx::Shape(%1355)\n  %1357 : Tensor = onnx::Shape(%1356)\n  %1358 : Tensor = onnx::ConstantOfShape[value={1}](%1357)\n  %1359 : Long() = onnx::Constant[value={-1}]()\n  %1360 : LongTensor = onnx::Mul(%1358, %1359)\n  %1361 : Tensor = onnx::Equal(%1356, %1360)\n  %1362 : Tensor = onnx::Where(%1361, %1358, %1356)\n  %1363 : Tensor = onnx::Expand(%1343, %1362)\n  %1364 : Tensor = onnx::Unsqueeze[axes=[-1]](%1363)\n  %1365 : Tensor = onnx::Shape(%1356)\n  %1366 : Tensor = onnx::ConstantOfShape[value={1}](%1365)\n  %1367 : Long() = onnx::Constant[value={-1}]()\n  %1368 : LongTensor = onnx::Mul(%1366, %1367)\n  %1369 : Tensor = onnx::Equal(%1356, %1368)\n  %1370 : Tensor = onnx::Where(%1369, %1366, %1356)\n  %1371 : Tensor = onnx::Expand(%1345, %1370)\n  %1372 : Tensor = onnx::Unsqueeze[axes=[-1]](%1371)\n  %1373 : Tensor = onnx::Shape(%1356)\n  %1374 : Tensor = onnx::ConstantOfShape[value={1}](%1373)\n  %1375 : Long() = onnx::Constant[value={-1}]()\n  %1376 : LongTensor = onnx::Mul(%1374, %1375)\n  %1377 : Tensor = onnx::Equal(%1356, %1376)\n  %1378 : Tensor = onnx::Where(%1377, %1374, %1356)\n  %1379 : Tensor = onnx::Expand(%1347, %1378)\n  %1380 : Tensor = onnx::Unsqueeze[axes=[-1]](%1379)\n  %1381 : Tensor = onnx::Shape(%1356)\n  %1382 : Tensor = onnx::ConstantOfShape[value={1}](%1381)\n  %1383 : Long() = onnx::Constant[value={-1}]()\n  %1384 : LongTensor = onnx::Mul(%1382, %1383)\n  %1385 : Tensor = onnx::Equal(%1356, %1384)\n  %1386 : Tensor = onnx::Where(%1385, %1382, %1356)\n  %1387 : Tensor = onnx::Expand(%1349, %1386)\n  %1388 : Tensor = onnx::Unsqueeze[axes=[-1]](%1387)\n  %1389 : Tensor = onnx::Shape(%1356)\n  %1390 : Tensor = onnx::ConstantOfShape[value={1}](%1389)\n  %1391 : Long() = onnx::Constant[value={-1}]()\n  %1392 : LongTensor = onnx::Mul(%1390, %1391)\n  %1393 : Tensor = onnx::Equal(%1356, %1392)\n  %1394 : Tensor = onnx::Where(%1393, %1390, %1356)\n  %1395 : Tensor = onnx::Expand(%1351, %1394)\n  %1396 : Tensor = onnx::Unsqueeze[axes=[-1]](%1395)\n  %1397 : Tensor = onnx::Concat[axis=-1](%1364, %1372, %1380, %1388, %1396)\n  %1398 : Tensor = onnx::Shape(%1282)\n  %1399 : Tensor = onnx::Constant[value={0}]()\n  %1400 : Tensor = onnx::Constant[value={5}]()\n  %1401 : Tensor = onnx::Constant[value={9223372036854775807}]()\n  %1402 : Tensor = onnx::Slice(%1398, %1400, %1401, %1399)\n  %1403 : Tensor = onnx::Concat[axis=0](%1356, %1402)\n  %1404 : Tensor = onnx::Reshape(%1301, %1403)\n  %1405 : Float(1:13440, 3:4480, 20:224, 32:7, 2:1) = onnx::ScatterND(%1282, %1397, %1404)\n  %1406 : Tensor = onnx::Constant[value={4}]()\n  %1407 : Tensor = onnx::Constant[value={0}]()\n  %1408 : Tensor = onnx::Constant[value={4}]()\n  %1409 : Tensor = onnx::Constant[value={1}]()\n  %1410 : Float(1:13440, 3:4480, 20:224, 32:7, 4:1) = onnx::Slice(%1405, %1407, %1408, %1406, %1409) # <ipython-input-6-cc76a8bf4ad5>:222:0\n  %1411 : Float() = onnx::Constant[value={16}]()\n  %1412 : Float(1:13440, 3:4480, 20:224, 32:7, 4:1) = onnx::Mul(%1410, %1411)\n  %1413 : Tensor = onnx::Constant[value=  3  20  32   4 [ CPULongType{4} ]]()\n  %1414 : Float(3:4480, 20:224, 32:7, 4:1) = onnx::Reshape(%1412, %1413) # <ipython-input-6-cc76a8bf4ad5>:222:0\n  %1415 : Tensor = onnx::Constant[value=  1   3  20  32   4 [ CPULongType{5} ]]()\n  %1417 : Tensor = onnx::ConstantOfShape[value={1}](%1556)\n  %1418 : Long() = onnx::Constant[value={-1}]()\n  %1419 : LongTensor = onnx::Mul(%1417, %1418)\n  %1420 : Long(5:1) = onnx::Constant[value=  1   3  20  32   4 [ CPULongType{5} ]]()\n  %1421 : Tensor = onnx::Equal(%1420, %1419)\n  %1422 : Tensor = onnx::Where(%1421, %1417, %1415)\n  %1423 : Float(1:13440, 3:4480, 20:224, 32:7, 4:1) = onnx::Expand(%1414, %1422) # <ipython-input-6-cc76a8bf4ad5>:222:0\n  %1424 : Tensor = onnx::Shape(%1405)\n  %1425 : Tensor = onnx::Constant[value={0}]()\n  %1426 : Long() = onnx::Gather[axis=0](%1424, %1425)\n  %1427 : Tensor = onnx::Cast[to=7](%1426)\n  %1428 : Tensor = onnx::Constant[value={0}]()\n  %1429 : Tensor = onnx::Constant[value={1}]()\n  %1430 : Tensor = onnx::Range(%1428, %1427, %1429)\n  %1431 : Tensor = onnx::Shape(%1405)\n  %1432 : Tensor = onnx::Constant[value={1}]()\n  %1433 : Long() = onnx::Gather[axis=0](%1431, %1432)\n  %1434 : Tensor = onnx::Cast[to=7](%1433)\n  %1435 : Tensor = onnx::Constant[value={0}]()\n  %1436 : Tensor = onnx::Constant[value={1}]()\n  %1437 : Tensor = onnx::Range(%1435, %1434, %1436)\n  %1438 : Tensor = onnx::Shape(%1405)\n  %1439 : Tensor = onnx::Constant[value={2}]()\n  %1440 : Long() = onnx::Gather[axis=0](%1438, %1439)\n  %1441 : Tensor = onnx::Cast[to=7](%1440)\n  %1442 : Tensor = onnx::Constant[value={0}]()\n  %1443 : Tensor = onnx::Constant[value={1}]()\n  %1444 : Tensor = onnx::Range(%1442, %1441, %1443)\n  %1445 : Tensor = onnx::Shape(%1405)\n  %1446 : Tensor = onnx::Constant[value={3}]()\n  %1447 : Long() = onnx::Gather[axis=0](%1445, %1446)\n  %1448 : Tensor = onnx::Cast[to=7](%1447)\n  %1449 : Tensor = onnx::Constant[value={0}]()\n  %1450 : Tensor = onnx::Constant[value={1}]()\n  %1451 : Tensor = onnx::Range(%1449, %1448, %1450)\n  %1452 : Tensor = onnx::Shape(%1405)\n  %1453 : Tensor = onnx::Constant[value={4}]()\n  %1454 : Long() = onnx::Gather[axis=0](%1452, %1453)\n  %1455 : Tensor = onnx::Cast[to=7](%1454)\n  %1456 : Tensor = onnx::Constant[value={0}]()\n  %1457 : Tensor = onnx::Constant[value={1}]()\n  %1458 : Tensor = onnx::Range(%1456, %1455, %1457)\n  %1459 : Tensor = onnx::Constant[value={0}]()\n  %1460 : Tensor = onnx::Constant[value={0}]()\n  %1461 : Tensor = onnx::Constant[value={4}]()\n  %1462 : Tensor = onnx::Constant[value={1}]()\n  %1463 : Tensor = onnx::Slice(%1458, %1460, %1461, %1459, %1462)\n  %1464 : Tensor = onnx::Constant[value=-1  1  1  1  1 [ CPULongType{5} ]]()\n  %1465 : Tensor = onnx::Reshape(%1430, %1464)\n  %1466 : Tensor = onnx::Constant[value=-1  1  1  1 [ CPULongType{4} ]]()\n  %1467 : Tensor = onnx::Reshape(%1437, %1466)\n  %1468 : Tensor = onnx::Constant[value=-1  1  1 [ CPULongType{3} ]]()\n  %1469 : Tensor = onnx::Reshape(%1444, %1468)\n  %1470 : Tensor = onnx::Constant[value=-1  1 [ CPULongType{2} ]]()\n  %1471 : Tensor = onnx::Reshape(%1451, %1470)\n  %1472 : Tensor = onnx::Constant[value={-1}]()\n  %1473 : Tensor = onnx::Reshape(%1463, %1472)\n  %1474 : Tensor = onnx::Add(%1465, %1467)\n  %1475 : Tensor = onnx::Add(%1474, %1469)\n  %1476 : Tensor = onnx::Add(%1475, %1471)\n  %1477 : Tensor = onnx::Add(%1476, %1473)\n  %1478 : Tensor = onnx::Shape(%1477)\n  %1479 : Tensor = onnx::Shape(%1478)\n  %1480 : Tensor = onnx::ConstantOfShape[value={1}](%1479)\n  %1481 : Long() = onnx::Constant[value={-1}]()\n  %1482 : LongTensor = onnx::Mul(%1480, %1481)\n  %1483 : Tensor = onnx::Equal(%1478, %1482)\n  %1484 : Tensor = onnx::Where(%1483, %1480, %1478)\n  %1485 : Tensor = onnx::Expand(%1465, %1484)\n  %1486 : Tensor = onnx::Unsqueeze[axes=[-1]](%1485)\n  %1487 : Tensor = onnx::Shape(%1478)\n  %1488 : Tensor = onnx::ConstantOfShape[value={1}](%1487)\n  %1489 : Long() = onnx::Constant[value={-1}]()\n  %1490 : LongTensor = onnx::Mul(%1488, %1489)\n  %1491 : Tensor = onnx::Equal(%1478, %1490)\n  %1492 : Tensor = onnx::Where(%1491, %1488, %1478)\n  %1493 : Tensor = onnx::Expand(%1467, %1492)\n  %1494 : Tensor = onnx::Unsqueeze[axes=[-1]](%1493)\n  %1495 : Tensor = onnx::Shape(%1478)\n  %1496 : Tensor = onnx::ConstantOfShape[value={1}](%1495)\n  %1497 : Long() = onnx::Constant[value={-1}]()\n  %1498 : LongTensor = onnx::Mul(%1496, %1497)\n  %1499 : Tensor = onnx::Equal(%1478, %1498)\n  %1500 : Tensor = onnx::Where(%1499, %1496, %1478)\n  %1501 : Tensor = onnx::Expand(%1469, %1500)\n  %1502 : Tensor = onnx::Unsqueeze[axes=[-1]](%1501)\n  %1503 : Tensor = onnx::Shape(%1478)\n  %1504 : Tensor = onnx::ConstantOfShape[value={1}](%1503)\n  %1505 : Long() = onnx::Constant[value={-1}]()\n  %1506 : LongTensor = onnx::Mul(%1504, %1505)\n  %1507 : Tensor = onnx::Equal(%1478, %1506)\n  %1508 : Tensor = onnx::Where(%1507, %1504, %1478)\n  %1509 : Tensor = onnx::Expand(%1471, %1508)\n  %1510 : Tensor = onnx::Unsqueeze[axes=[-1]](%1509)\n  %1511 : Tensor = onnx::Shape(%1478)\n  %1512 : Tensor = onnx::ConstantOfShape[value={1}](%1511)\n  %1513 : Long() = onnx::Constant[value={-1}]()\n  %1514 : LongTensor = onnx::Mul(%1512, %1513)\n  %1515 : Tensor = onnx::Equal(%1478, %1514)\n  %1516 : Tensor = onnx::Where(%1515, %1512, %1478)\n  %1517 : Tensor = onnx::Expand(%1473, %1516)\n  %1518 : Tensor = onnx::Unsqueeze[axes=[-1]](%1517)\n  %1519 : Tensor = onnx::Concat[axis=-1](%1486, %1494, %1502, %1510, %1518)\n  %1520 : Tensor = onnx::Shape(%1405)\n  %1521 : Tensor = onnx::Constant[value={0}]()\n  %1522 : Tensor = onnx::Constant[value={5}]()\n  %1523 : Tensor = onnx::Constant[value={9223372036854775807}]()\n  %1524 : Tensor = onnx::Slice(%1520, %1522, %1523, %1521)\n  %1525 : Tensor = onnx::Concat[axis=0](%1478, %1524)\n  %1526 : Tensor = onnx::Reshape(%1423, %1525)\n  %1527 : Float(1:13440, 3:4480, 20:224, 32:7, 4:1) = onnx::ScatterND(%1405, %1519, %1526)\n  %1530 : Tensor = onnx::Unsqueeze[axes=[0]](%1147)\n  %1533 : Tensor = onnx::Concat[axis=0](%1530, %1557, %1558)\n  %1534 : Float(1:13440, 1920:7, 7:1) = onnx::Reshape(%1527, %1533) # <ipython-input-6-cc76a8bf4ad5>:224:0\n  %classes : Float(1:16800, 2400:7, 7:1) = onnx::Concat[axis=1](%1119, %1534) # <ipython-input-6-cc76a8bf4ad5>:314:0\n  return (%classes, %yolo1, %yolo2)\n\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Export the model\n",
    "torch.onnx.export(model,                     # model being run\n",
    "                  x,                         # model input (or a tuple for multiple inputs)\n",
    "                  \"weights/yolo-fastest.onnx\",   # where to save the model (can be a file or file-like object)\n",
    "                  export_params=True,        # store the trained parameter weights inside the model file\n",
    "                  opset_version=11,          # the ONNX version to export the model to\n",
    "                  do_constant_folding=True,  # whether to execute constant folding for optimization\n",
    "                  verbose=True,\n",
    "                  input_names = ['images'],   # the model's input names\n",
    "                  output_names = ['classes','yolo1','yolo2'], # the model's output names\n",
    "                  dynamic_axes={'input' : {0 : 'batch_size'},    # variable lenght axes\n",
    "                                'output' : {0 : 'batch_size'}}\n",
    "                 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "import onnx\n",
    "\n",
    "onnx_model = onnx.load(\"weights/yolo-fastest.onnx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "tags": [
     "outputPrepend"
    ]
   },
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "= <Scalar Tensor []>]()\n  %1029 = Range(%1027, %1026, %1028)\n  %1030 = Shape(%990)\n  %1031 = Constant[value = <Scalar Tensor []>]()\n  %1032 = Gather[axis = 0](%1030, %1031)\n  %1033 = Cast[to = 7](%1032)\n  %1034 = Constant[value = <Scalar Tensor []>]()\n  %1035 = Constant[value = <Scalar Tensor []>]()\n  %1036 = Range(%1034, %1033, %1035)\n  %1037 = Shape(%990)\n  %1038 = Constant[value = <Scalar Tensor []>]()\n  %1039 = Gather[axis = 0](%1037, %1038)\n  %1040 = Cast[to = 7](%1039)\n  %1041 = Constant[value = <Scalar Tensor []>]()\n  %1042 = Constant[value = <Scalar Tensor []>]()\n  %1043 = Range(%1041, %1040, %1042)\n  %1044 = Constant[value = <Tensor>]()\n  %1045 = Constant[value = <Tensor>]()\n  %1046 = Constant[value = <Tensor>]()\n  %1047 = Constant[value = <Tensor>]()\n  %1048 = Slice(%1043, %1045, %1046, %1044, %1047)\n  %1049 = Constant[value = <Tensor>]()\n  %1050 = Reshape(%1015, %1049)\n  %1051 = Constant[value = <Tensor>]()\n  %1052 = Reshape(%1022, %1051)\n  %1053 = Constant[value = <Tensor>]()\n  %1054 = Reshape(%1029, %1053)\n  %1055 = Constant[value = <Tensor>]()\n  %1056 = Reshape(%1036, %1055)\n  %1057 = Constant[value = <Tensor>]()\n  %1058 = Reshape(%1048, %1057)\n  %1059 = Add(%1050, %1052)\n  %1060 = Add(%1059, %1054)\n  %1061 = Add(%1060, %1056)\n  %1062 = Add(%1061, %1058)\n  %1063 = Shape(%1062)\n  %1064 = Shape(%1063)\n  %1065 = ConstantOfShape[value = <Tensor>](%1064)\n  %1066 = Constant[value = <Scalar Tensor []>]()\n  %1067 = Mul(%1065, %1066)\n  %1068 = Equal(%1063, %1067)\n  %1069 = Where(%1068, %1065, %1063)\n  %1070 = Expand(%1050, %1069)\n  %1071 = Unsqueeze[axes = [-1]](%1070)\n  %1072 = Shape(%1063)\n  %1073 = ConstantOfShape[value = <Tensor>](%1072)\n  %1074 = Constant[value = <Scalar Tensor []>]()\n  %1075 = Mul(%1073, %1074)\n  %1076 = Equal(%1063, %1075)\n  %1077 = Where(%1076, %1073, %1063)\n  %1078 = Expand(%1052, %1077)\n  %1079 = Unsqueeze[axes = [-1]](%1078)\n  %1080 = Shape(%1063)\n  %1081 = ConstantOfShape[value = <Tensor>](%1080)\n  %1082 = Constant[value = <Scalar Tensor []>]()\n  %1083 = Mul(%1081, %1082)\n  %1084 = Equal(%1063, %1083)\n  %1085 = Where(%1084, %1081, %1063)\n  %1086 = Expand(%1054, %1085)\n  %1087 = Unsqueeze[axes = [-1]](%1086)\n  %1088 = Shape(%1063)\n  %1089 = ConstantOfShape[value = <Tensor>](%1088)\n  %1090 = Constant[value = <Scalar Tensor []>]()\n  %1091 = Mul(%1089, %1090)\n  %1092 = Equal(%1063, %1091)\n  %1093 = Where(%1092, %1089, %1063)\n  %1094 = Expand(%1056, %1093)\n  %1095 = Unsqueeze[axes = [-1]](%1094)\n  %1096 = Shape(%1063)\n  %1097 = ConstantOfShape[value = <Tensor>](%1096)\n  %1098 = Constant[value = <Scalar Tensor []>]()\n  %1099 = Mul(%1097, %1098)\n  %1100 = Equal(%1063, %1099)\n  %1101 = Where(%1100, %1097, %1063)\n  %1102 = Expand(%1058, %1101)\n  %1103 = Unsqueeze[axes = [-1]](%1102)\n  %1104 = Concat[axis = -1](%1071, %1079, %1087, %1095, %1103)\n  %1105 = Shape(%990)\n  %1106 = Constant[value = <Tensor>]()\n  %1107 = Constant[value = <Tensor>]()\n  %1108 = Constant[value = <Tensor>]()\n  %1109 = Slice(%1105, %1107, %1108, %1106)\n  %1110 = Concat[axis = 0](%1063, %1109)\n  %1111 = Reshape(%1008, %1110)\n  %1112 = ScatterND(%990, %1104, %1111)\n  %1115 = Unsqueeze[axes = [0]](%732)\n  %1118 = Concat[axis = 0](%1115, %1543, %1544)\n  %1119 = Reshape(%1112, %1118)\n  %1128 = Constant[value = <Tensor>]()\n  %1129 = Resize[coordinate_transformation_mode = 'asymmetric', cubic_coeff_a = -0.75, mode = 'nearest', nearest_mode = 'floor'](%718, %1128, %1549)\n  %1130 = Concat[axis = 1](%1129, %665)\n  %1131 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1130, %module_list.119.Conv2d.weight)\n  %1132 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1131, %module_list.119.BatchNorm2d.weight, %module_list.119.BatchNorm2d.bias, %module_list.119.BatchNorm2d.running_mean, %module_list.119.BatchNorm2d.running_var)\n  %1133 = LeakyRelu[alpha = 0.100000001490116](%1132)\n  %1134 = Conv[dilations = [1, 1], group = 96, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%1133, %module_list.120.Conv2d.weight)\n  %1135 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1134, %module_list.120.BatchNorm2d.weight, %module_list.120.BatchNorm2d.bias, %module_list.120.BatchNorm2d.running_mean, %module_list.120.BatchNorm2d.running_var)\n  %1136 = LeakyRelu[alpha = 0.100000001490116](%1135)\n  %1137 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1136, %module_list.121.Conv2d.weight)\n  %1138 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1137, %module_list.121.BatchNorm2d.weight, %module_list.121.BatchNorm2d.bias, %module_list.121.BatchNorm2d.running_mean, %module_list.121.BatchNorm2d.running_var)\n  %1139 = Conv[dilations = [1, 1], group = 96, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%1138, %module_list.122.Conv2d.weight)\n  %1140 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1139, %module_list.122.BatchNorm2d.weight, %module_list.122.BatchNorm2d.bias, %module_list.122.BatchNorm2d.running_mean, %module_list.122.BatchNorm2d.running_var)\n  %1141 = LeakyRelu[alpha = 0.100000001490116](%1140)\n  %1142 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1141, %module_list.123.Conv2d.weight)\n  %1143 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1142, %module_list.123.BatchNorm2d.weight, %module_list.123.BatchNorm2d.bias, %module_list.123.BatchNorm2d.running_mean, %module_list.123.BatchNorm2d.running_var)\n  %1144 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1143, %module_list.124.Conv2d.weight, %module_list.124.Conv2d.bias)\n  %1145 = Shape(%1144)\n  %1146 = Constant[value = <Scalar Tensor []>]()\n  %1147 = Gather[axis = 0](%1145, %1146)\n  %1152 = Unsqueeze[axes = [0]](%1147)\n  %1157 = Concat[axis = 0](%1152, %1550, %1551, %1552, %1553)\n  %1158 = Reshape(%1144, %1157)\n  %yolo2 = Transpose[perm = [0, 1, 3, 4, 2]](%1158)\n  %1160 = Constant[value = <Tensor>]()\n  %1161 = Constant[value = <Tensor>]()\n  %1162 = Constant[value = <Tensor>]()\n  %1163 = Constant[value = <Tensor>]()\n  %1164 = Slice(%yolo2, %1161, %1162, %1160, %1163)\n  %1165 = Sigmoid(%1164)\n  %1166 = Constant[value = <Tensor>]()\n  %1167 = Add(%1165, %1166)\n  %1168 = Constant[value = <Tensor>]()\n  %1169 = Reshape(%1167, %1168)\n  %1170 = Constant[value = <Tensor>]()\n  %1172 = ConstantOfShape[value = <Tensor>](%1554)\n  %1173 = Constant[value = <Scalar Tensor []>]()\n  %1174 = Mul(%1172, %1173)\n  %1175 = Constant[value = <Tensor>]()\n  %1176 = Equal(%1175, %1174)\n  %1177 = Where(%1176, %1172, %1170)\n  %1178 = Expand(%1169, %1177)\n  %1179 = Shape(%yolo2)\n  %1180 = Constant[value = <Scalar Tensor []>]()\n  %1181 = Gather[axis = 0](%1179, %1180)\n  %1182 = Cast[to = 7](%1181)\n  %1183 = Constant[value = <Scalar Tensor []>]()\n  %1184 = Constant[value = <Scalar Tensor []>]()\n  %1185 = Range(%1183, %1182, %1184)\n  %1186 = Shape(%yolo2)\n  %1187 = Constant[value = <Scalar Tensor []>]()\n  %1188 = Gather[axis = 0](%1186, %1187)\n  %1189 = Cast[to = 7](%1188)\n  %1190 = Constant[value = <Scalar Tensor []>]()\n  %1191 = Constant[value = <Scalar Tensor []>]()\n  %1192 = Range(%1190, %1189, %1191)\n  %1193 = Shape(%yolo2)\n  %1194 = Constant[value = <Scalar Tensor []>]()\n  %1195 = Gather[axis = 0](%1193, %1194)\n  %1196 = Cast[to = 7](%1195)\n  %1197 = Constant[value = <Scalar Tensor []>]()\n  %1198 = Constant[value = <Scalar Tensor []>]()\n  %1199 = Range(%1197, %1196, %1198)\n  %1200 = Shape(%yolo2)\n  %1201 = Constant[value = <Scalar Tensor []>]()\n  %1202 = Gather[axis = 0](%1200, %1201)\n  %1203 = Cast[to = 7](%1202)\n  %1204 = Constant[value = <Scalar Tensor []>]()\n  %1205 = Constant[value = <Scalar Tensor []>]()\n  %1206 = Range(%1204, %1203, %1205)\n  %1207 = Shape(%yolo2)\n  %1208 = Constant[value = <Scalar Tensor []>]()\n  %1209 = Gather[axis = 0](%1207, %1208)\n  %1210 = Cast[to = 7](%1209)\n  %1211 = Constant[value = <Scalar Tensor []>]()\n  %1212 = Constant[value = <Scalar Tensor []>]()\n  %1213 = Range(%1211, %1210, %1212)\n  %1214 = Constant[value = <Tensor>]()\n  %1215 = Constant[value = <Tensor>]()\n  %1216 = Constant[value = <Tensor>]()\n  %1217 = Constant[value = <Tensor>]()\n  %1218 = Slice(%1213, %1215, %1216, %1214, %1217)\n  %1219 = Constant[value = <Tensor>]()\n  %1220 = Reshape(%1185, %1219)\n  %1221 = Constant[value = <Tensor>]()\n  %1222 = Reshape(%1192, %1221)\n  %1223 = Constant[value = <Tensor>]()\n  %1224 = Reshape(%1199, %1223)\n  %1225 = Constant[value = <Tensor>]()\n  %1226 = Reshape(%1206, %1225)\n  %1227 = Constant[value = <Tensor>]()\n  %1228 = Reshape(%1218, %1227)\n  %1229 = Add(%1220, %1222)\n  %1230 = Add(%1229, %1224)\n  %1231 = Add(%1230, %1226)\n  %1232 = Add(%1231, %1228)\n  %1233 = Shape(%1232)\n  %1234 = Shape(%1233)\n  %1235 = ConstantOfShape[value = <Tensor>](%1234)\n  %1236 = Constant[value = <Scalar Tensor []>]()\n  %1237 = Mul(%1235, %1236)\n  %1238 = Equal(%1233, %1237)\n  %1239 = Where(%1238, %1235, %1233)\n  %1240 = Expand(%1220, %1239)\n  %1241 = Unsqueeze[axes = [-1]](%1240)\n  %1242 = Shape(%1233)\n  %1243 = ConstantOfShape[value = <Tensor>](%1242)\n  %1244 = Constant[value = <Scalar Tensor []>]()\n  %1245 = Mul(%1243, %1244)\n  %1246 = Equal(%1233, %1245)\n  %1247 = Where(%1246, %1243, %1233)\n  %1248 = Expand(%1222, %1247)\n  %1249 = Unsqueeze[axes = [-1]](%1248)\n  %1250 = Shape(%1233)\n  %1251 = ConstantOfShape[value = <Tensor>](%1250)\n  %1252 = Constant[value = <Scalar Tensor []>]()\n  %1253 = Mul(%1251, %1252)\n  %1254 = Equal(%1233, %1253)\n  %1255 = Where(%1254, %1251, %1233)\n  %1256 = Expand(%1224, %1255)\n  %1257 = Unsqueeze[axes = [-1]](%1256)\n  %1258 = Shape(%1233)\n  %1259 = ConstantOfShape[value = <Tensor>](%1258)\n  %1260 = Constant[value = <Scalar Tensor []>]()\n  %1261 = Mul(%1259, %1260)\n  %1262 = Equal(%1233, %1261)\n  %1263 = Where(%1262, %1259, %1233)\n  %1264 = Expand(%1226, %1263)\n  %1265 = Unsqueeze[axes = [-1]](%1264)\n  %1266 = Shape(%1233)\n  %1267 = ConstantOfShape[value = <Tensor>](%1266)\n  %1268 = Constant[value = <Scalar Tensor []>]()\n  %1269 = Mul(%1267, %1268)\n  %1270 = Equal(%1233, %1269)\n  %1271 = Where(%1270, %1267, %1233)\n  %1272 = Expand(%1228, %1271)\n  %1273 = Unsqueeze[axes = [-1]](%1272)\n  %1274 = Concat[axis = -1](%1241, %1249, %1257, %1265, %1273)\n  %1275 = Shape(%yolo2)\n  %1276 = Constant[value = <Tensor>]()\n  %1277 = Constant[value = <Tensor>]()\n  %1278 = Constant[value = <Tensor>]()\n  %1279 = Slice(%1275, %1277, %1278, %1276)\n  %1280 = Concat[axis = 0](%1233, %1279)\n  %1281 = Reshape(%1178, %1280)\n  %1282 = ScatterND(%yolo2, %1274, %1281)\n  %1283 = Constant[value = <Tensor>]()\n  %1284 = Constant[value = <Tensor>]()\n  %1285 = Constant[value = <Tensor>]()\n  %1286 = Constant[value = <Tensor>]()\n  %1287 = Slice(%1282, %1284, %1285, %1283, %1286)\n  %1288 = Exp(%1287)\n  %1289 = Constant[value = <Tensor>]()\n  %1290 = Mul(%1288, %1289)\n  %1291 = Constant[value = <Tensor>]()\n  %1292 = Reshape(%1290, %1291)\n  %1293 = Constant[value = <Tensor>]()\n  %1295 = ConstantOfShape[value = <Tensor>](%1555)\n  %1296 = Constant[value = <Scalar Tensor []>]()\n  %1297 = Mul(%1295, %1296)\n  %1298 = Constant[value = <Tensor>]()\n  %1299 = Equal(%1298, %1297)\n  %1300 = Where(%1299, %1295, %1293)\n  %1301 = Expand(%1292, %1300)\n  %1302 = Shape(%1282)\n  %1303 = Constant[value = <Scalar Tensor []>]()\n  %1304 = Gather[axis = 0](%1302, %1303)\n  %1305 = Cast[to = 7](%1304)\n  %1306 = Constant[value = <Scalar Tensor []>]()\n  %1307 = Constant[value = <Scalar Tensor []>]()\n  %1308 = Range(%1306, %1305, %1307)\n  %1309 = Shape(%1282)\n  %1310 = Constant[value = <Scalar Tensor []>]()\n  %1311 = Gather[axis = 0](%1309, %1310)\n  %1312 = Cast[to = 7](%1311)\n  %1313 = Constant[value = <Scalar Tensor []>]()\n  %1314 = Constant[value = <Scalar Tensor []>]()\n  %1315 = Range(%1313, %1312, %1314)\n  %1316 = Shape(%1282)\n  %1317 = Constant[value = <Scalar Tensor []>]()\n  %1318 = Gather[axis = 0](%1316, %1317)\n  %1319 = Cast[to = 7](%1318)\n  %1320 = Constant[value = <Scalar Tensor []>]()\n  %1321 = Constant[value = <Scalar Tensor []>]()\n  %1322 = Range(%1320, %1319, %1321)\n  %1323 = Shape(%1282)\n  %1324 = Constant[value = <Scalar Tensor []>]()\n  %1325 = Gather[axis = 0](%1323, %1324)\n  %1326 = Cast[to = 7](%1325)\n  %1327 = Constant[value = <Scalar Tensor []>]()\n  %1328 = Constant[value = <Scalar Tensor []>]()\n  %1329 = Range(%1327, %1326, %1328)\n  %1330 = Shape(%1282)\n  %1331 = Constant[value = <Scalar Tensor []>]()\n  %1332 = Gather[axis = 0](%1330, %1331)\n  %1333 = Cast[to = 7](%1332)\n  %1334 = Constant[value = <Scalar Tensor []>]()\n  %1335 = Constant[value = <Scalar Tensor []>]()\n  %1336 = Range(%1334, %1333, %1335)\n  %1337 = Constant[value = <Tensor>]()\n  %1338 = Constant[value = <Tensor>]()\n  %1339 = Constant[value = <Tensor>]()\n  %1340 = Constant[value = <Tensor>]()\n  %1341 = Slice(%1336, %1338, %1339, %1337, %1340)\n  %1342 = Constant[value = <Tensor>]()\n  %1343 = Reshape(%1308, %1342)\n  %1344 = Constant[value = <Tensor>]()\n  %1345 = Reshape(%1315, %1344)\n  %1346 = Constant[value = <Tensor>]()\n  %1347 = Reshape(%1322, %1346)\n  %1348 = Constant[value = <Tensor>]()\n  %1349 = Reshape(%1329, %1348)\n  %1350 = Constant[value = <Tensor>]()\n  %1351 = Reshape(%1341, %1350)\n  %1352 = Add(%1343, %1345)\n  %1353 = Add(%1352, %1347)\n  %1354 = Add(%1353, %1349)\n  %1355 = Add(%1354, %1351)\n  %1356 = Shape(%1355)\n  %1357 = Shape(%1356)\n  %1358 = ConstantOfShape[value = <Tensor>](%1357)\n  %1359 = Constant[value = <Scalar Tensor []>]()\n  %1360 = Mul(%1358, %1359)\n  %1361 = Equal(%1356, %1360)\n  %1362 = Where(%1361, %1358, %1356)\n  %1363 = Expand(%1343, %1362)\n  %1364 = Unsqueeze[axes = [-1]](%1363)\n  %1365 = Shape(%1356)\n  %1366 = ConstantOfShape[value = <Tensor>](%1365)\n  %1367 = Constant[value = <Scalar Tensor []>]()\n  %1368 = Mul(%1366, %1367)\n  %1369 = Equal(%1356, %1368)\n  %1370 = Where(%1369, %1366, %1356)\n  %1371 = Expand(%1345, %1370)\n  %1372 = Unsqueeze[axes = [-1]](%1371)\n  %1373 = Shape(%1356)\n  %1374 = ConstantOfShape[value = <Tensor>](%1373)\n  %1375 = Constant[value = <Scalar Tensor []>]()\n  %1376 = Mul(%1374, %1375)\n  %1377 = Equal(%1356, %1376)\n  %1378 = Where(%1377, %1374, %1356)\n  %1379 = Expand(%1347, %1378)\n  %1380 = Unsqueeze[axes = [-1]](%1379)\n  %1381 = Shape(%1356)\n  %1382 = ConstantOfShape[value = <Tensor>](%1381)\n  %1383 = Constant[value = <Scalar Tensor []>]()\n  %1384 = Mul(%1382, %1383)\n  %1385 = Equal(%1356, %1384)\n  %1386 = Where(%1385, %1382, %1356)\n  %1387 = Expand(%1349, %1386)\n  %1388 = Unsqueeze[axes = [-1]](%1387)\n  %1389 = Shape(%1356)\n  %1390 = ConstantOfShape[value = <Tensor>](%1389)\n  %1391 = Constant[value = <Scalar Tensor []>]()\n  %1392 = Mul(%1390, %1391)\n  %1393 = Equal(%1356, %1392)\n  %1394 = Where(%1393, %1390, %1356)\n  %1395 = Expand(%1351, %1394)\n  %1396 = Unsqueeze[axes = [-1]](%1395)\n  %1397 = Concat[axis = -1](%1364, %1372, %1380, %1388, %1396)\n  %1398 = Shape(%1282)\n  %1399 = Constant[value = <Tensor>]()\n  %1400 = Constant[value = <Tensor>]()\n  %1401 = Constant[value = <Tensor>]()\n  %1402 = Slice(%1398, %1400, %1401, %1399)\n  %1403 = Concat[axis = 0](%1356, %1402)\n  %1404 = Reshape(%1301, %1403)\n  %1405 = ScatterND(%1282, %1397, %1404)\n  %1406 = Constant[value = <Tensor>]()\n  %1407 = Constant[value = <Tensor>]()\n  %1408 = Constant[value = <Tensor>]()\n  %1409 = Constant[value = <Tensor>]()\n  %1410 = Slice(%1405, %1407, %1408, %1406, %1409)\n  %1411 = Constant[value = <Scalar Tensor []>]()\n  %1412 = Mul(%1410, %1411)\n  %1413 = Constant[value = <Tensor>]()\n  %1414 = Reshape(%1412, %1413)\n  %1415 = Constant[value = <Tensor>]()\n  %1417 = ConstantOfShape[value = <Tensor>](%1556)\n  %1418 = Constant[value = <Scalar Tensor []>]()\n  %1419 = Mul(%1417, %1418)\n  %1420 = Constant[value = <Tensor>]()\n  %1421 = Equal(%1420, %1419)\n  %1422 = Where(%1421, %1417, %1415)\n  %1423 = Expand(%1414, %1422)\n  %1424 = Shape(%1405)\n  %1425 = Constant[value = <Scalar Tensor []>]()\n  %1426 = Gather[axis = 0](%1424, %1425)\n  %1427 = Cast[to = 7](%1426)\n  %1428 = Constant[value = <Scalar Tensor []>]()\n  %1429 = Constant[value = <Scalar Tensor []>]()\n  %1430 = Range(%1428, %1427, %1429)\n  %1431 = Shape(%1405)\n  %1432 = Constant[value = <Scalar Tensor []>]()\n  %1433 = Gather[axis = 0](%1431, %1432)\n  %1434 = Cast[to = 7](%1433)\n  %1435 = Constant[value = <Scalar Tensor []>]()\n  %1436 = Constant[value = <Scalar Tensor []>]()\n  %1437 = Range(%1435, %1434, %1436)\n  %1438 = Shape(%1405)\n  %1439 = Constant[value = <Scalar Tensor []>]()\n  %1440 = Gather[axis = 0](%1438, %1439)\n  %1441 = Cast[to = 7](%1440)\n  %1442 = Constant[value = <Scalar Tensor []>]()\n  %1443 = Constant[value = <Scalar Tensor []>]()\n  %1444 = Range(%1442, %1441, %1443)\n  %1445 = Shape(%1405)\n  %1446 = Constant[value = <Scalar Tensor []>]()\n  %1447 = Gather[axis = 0](%1445, %1446)\n  %1448 = Cast[to = 7](%1447)\n  %1449 = Constant[value = <Scalar Tensor []>]()\n  %1450 = Constant[value = <Scalar Tensor []>]()\n  %1451 = Range(%1449, %1448, %1450)\n  %1452 = Shape(%1405)\n  %1453 = Constant[value = <Scalar Tensor []>]()\n  %1454 = Gather[axis = 0](%1452, %1453)\n  %1455 = Cast[to = 7](%1454)\n  %1456 = Constant[value = <Scalar Tensor []>]()\n  %1457 = Constant[value = <Scalar Tensor []>]()\n  %1458 = Range(%1456, %1455, %1457)\n  %1459 = Constant[value = <Tensor>]()\n  %1460 = Constant[value = <Tensor>]()\n  %1461 = Constant[value = <Tensor>]()\n  %1462 = Constant[value = <Tensor>]()\n  %1463 = Slice(%1458, %1460, %1461, %1459, %1462)\n  %1464 = Constant[value = <Tensor>]()\n  %1465 = Reshape(%1430, %1464)\n  %1466 = Constant[value = <Tensor>]()\n  %1467 = Reshape(%1437, %1466)\n  %1468 = Constant[value = <Tensor>]()\n  %1469 = Reshape(%1444, %1468)\n  %1470 = Constant[value = <Tensor>]()\n  %1471 = Reshape(%1451, %1470)\n  %1472 = Constant[value = <Tensor>]()\n  %1473 = Reshape(%1463, %1472)\n  %1474 = Add(%1465, %1467)\n  %1475 = Add(%1474, %1469)\n  %1476 = Add(%1475, %1471)\n  %1477 = Add(%1476, %1473)\n  %1478 = Shape(%1477)\n  %1479 = Shape(%1478)\n  %1480 = ConstantOfShape[value = <Tensor>](%1479)\n  %1481 = Constant[value = <Scalar Tensor []>]()\n  %1482 = Mul(%1480, %1481)\n  %1483 = Equal(%1478, %1482)\n  %1484 = Where(%1483, %1480, %1478)\n  %1485 = Expand(%1465, %1484)\n  %1486 = Unsqueeze[axes = [-1]](%1485)\n  %1487 = Shape(%1478)\n  %1488 = ConstantOfShape[value = <Tensor>](%1487)\n  %1489 = Constant[value = <Scalar Tensor []>]()\n  %1490 = Mul(%1488, %1489)\n  %1491 = Equal(%1478, %1490)\n  %1492 = Where(%1491, %1488, %1478)\n  %1493 = Expand(%1467, %1492)\n  %1494 = Unsqueeze[axes = [-1]](%1493)\n  %1495 = Shape(%1478)\n  %1496 = ConstantOfShape[value = <Tensor>](%1495)\n  %1497 = Constant[value = <Scalar Tensor []>]()\n  %1498 = Mul(%1496, %1497)\n  %1499 = Equal(%1478, %1498)\n  %1500 = Where(%1499, %1496, %1478)\n  %1501 = Expand(%1469, %1500)\n  %1502 = Unsqueeze[axes = [-1]](%1501)\n  %1503 = Shape(%1478)\n  %1504 = ConstantOfShape[value = <Tensor>](%1503)\n  %1505 = Constant[value = <Scalar Tensor []>]()\n  %1506 = Mul(%1504, %1505)\n  %1507 = Equal(%1478, %1506)\n  %1508 = Where(%1507, %1504, %1478)\n  %1509 = Expand(%1471, %1508)\n  %1510 = Unsqueeze[axes = [-1]](%1509)\n  %1511 = Shape(%1478)\n  %1512 = ConstantOfShape[value = <Tensor>](%1511)\n  %1513 = Constant[value = <Scalar Tensor []>]()\n  %1514 = Mul(%1512, %1513)\n  %1515 = Equal(%1478, %1514)\n  %1516 = Where(%1515, %1512, %1478)\n  %1517 = Expand(%1473, %1516)\n  %1518 = Unsqueeze[axes = [-1]](%1517)\n  %1519 = Concat[axis = -1](%1486, %1494, %1502, %1510, %1518)\n  %1520 = Shape(%1405)\n  %1521 = Constant[value = <Tensor>]()\n  %1522 = Constant[value = <Tensor>]()\n  %1523 = Constant[value = <Tensor>]()\n  %1524 = Slice(%1520, %1522, %1523, %1521)\n  %1525 = Concat[axis = 0](%1478, %1524)\n  %1526 = Reshape(%1423, %1525)\n  %1527 = ScatterND(%1405, %1519, %1526)\n  %1530 = Unsqueeze[axes = [0]](%1147)\n  %1533 = Concat[axis = 0](%1530, %1557, %1558)\n  %1534 = Reshape(%1527, %1533)\n  %classes = Concat[axis = 1](%1119, %1534)\n  return %classes, %yolo1, %yolo2\n}\n"
     ]
    }
   ],
   "source": [
    "onnx.checker.check_model(onnx_model)  # Check that the IR is well formed        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "graph torch-jit-export (\n  %images[FLOAT, 1x3x320x512]\n) initializers (\n  %1536[INT64, 1]\n  %1537[INT64, 1]\n  %1538[INT64, 1]\n  %1539[INT64, 1]\n  %1540[INT64, 1]\n  %1541[INT64, 1]\n  %1542[INT64, 1]\n  %1543[INT64, 1]\n  %1544[INT64, 1]\n  %1549[FLOAT, 4]\n  %1550[INT64, 1]\n  %1551[INT64, 1]\n  %1552[INT64, 1]\n  %1553[INT64, 1]\n  %1554[INT64, 1]\n  %1555[INT64, 1]\n  %1556[INT64, 1]\n  %1557[INT64, 1]\n  %1558[INT64, 1]\n  %module_list.0.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.0.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.0.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.0.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.0.Conv2d.weight[FLOAT, 8x3x3x3]\n  %module_list.1.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.1.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.1.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.1.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.1.Conv2d.weight[FLOAT, 8x8x1x1]\n  %module_list.10.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.10.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.10.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.10.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.10.Conv2d.weight[FLOAT, 24x1x3x3]\n  %module_list.100.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.100.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.100.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.100.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.100.Conv2d.weight[FLOAT, 224x1x3x3]\n  %module_list.101.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.101.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.101.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.101.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.101.Conv2d.weight[FLOAT, 48x224x1x1]\n  %module_list.104.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.104.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.104.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.104.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.104.Conv2d.weight[FLOAT, 224x48x1x1]\n  %module_list.105.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.105.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.105.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.105.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.105.Conv2d.weight[FLOAT, 224x1x3x3]\n  %module_list.106.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.106.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.106.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.106.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.106.Conv2d.weight[FLOAT, 48x224x1x1]\n  %module_list.109.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.109.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.109.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.109.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.109.Conv2d.weight[FLOAT, 96x48x1x1]\n  %module_list.11.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.11.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.11.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.11.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.11.Conv2d.weight[FLOAT, 8x24x1x1]\n  %module_list.110.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.110.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.110.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.110.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.110.Conv2d.weight[FLOAT, 96x1x5x5]\n  %module_list.111.BatchNorm2d.bias[FLOAT, 128]\n  %module_list.111.BatchNorm2d.running_mean[FLOAT, 128]\n  %module_list.111.BatchNorm2d.running_var[FLOAT, 128]\n  %module_list.111.BatchNorm2d.weight[FLOAT, 128]\n  %module_list.111.Conv2d.weight[FLOAT, 128x96x1x1]\n  %module_list.112.BatchNorm2d.bias[FLOAT, 128]\n  %module_list.112.BatchNorm2d.running_mean[FLOAT, 128]\n  %module_list.112.BatchNorm2d.running_var[FLOAT, 128]\n  %module_list.112.BatchNorm2d.weight[FLOAT, 128]\n  %module_list.112.Conv2d.weight[FLOAT, 128x1x5x5]\n  %module_list.113.BatchNorm2d.bias[FLOAT, 128]\n  %module_list.113.BatchNorm2d.running_mean[FLOAT, 128]\n  %module_list.113.BatchNorm2d.running_var[FLOAT, 128]\n  %module_list.113.BatchNorm2d.weight[FLOAT, 128]\n  %module_list.113.Conv2d.weight[FLOAT, 128x128x1x1]\n  %module_list.114.Conv2d.bias[FLOAT, 21]\n  %module_list.114.Conv2d.weight[FLOAT, 21x128x1x1]\n  %module_list.119.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.119.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.119.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.119.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.119.Conv2d.weight[FLOAT, 96x232x1x1]\n  %module_list.12.BatchNorm2d.bias[FLOAT, 32]\n  %module_list.12.BatchNorm2d.running_mean[FLOAT, 32]\n  %module_list.12.BatchNorm2d.running_var[FLOAT, 32]\n  %module_list.12.BatchNorm2d.weight[FLOAT, 32]\n  %module_list.12.Conv2d.weight[FLOAT, 32x8x1x1]\n  %module_list.120.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.120.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.120.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.120.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.120.Conv2d.weight[FLOAT, 96x1x5x5]\n  %module_list.121.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.121.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.121.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.121.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.121.Conv2d.weight[FLOAT, 96x96x1x1]\n  %module_list.122.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.122.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.122.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.122.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.122.Conv2d.weight[FLOAT, 96x1x5x5]\n  %module_list.123.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.123.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.123.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.123.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.123.Conv2d.weight[FLOAT, 96x96x1x1]\n  %module_list.124.Conv2d.bias[FLOAT, 21]\n  %module_list.124.Conv2d.weight[FLOAT, 21x96x1x1]\n  %module_list.13.BatchNorm2d.bias[FLOAT, 32]\n  %module_list.13.BatchNorm2d.running_mean[FLOAT, 32]\n  %module_list.13.BatchNorm2d.running_var[FLOAT, 32]\n  %module_list.13.BatchNorm2d.weight[FLOAT, 32]\n  %module_list.13.Conv2d.weight[FLOAT, 32x1x3x3]\n  %module_list.14.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.14.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.14.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.14.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.14.Conv2d.weight[FLOAT, 8x32x1x1]\n  %module_list.17.BatchNorm2d.bias[FLOAT, 32]\n  %module_list.17.BatchNorm2d.running_mean[FLOAT, 32]\n  %module_list.17.BatchNorm2d.running_var[FLOAT, 32]\n  %module_list.17.BatchNorm2d.weight[FLOAT, 32]\n  %module_list.17.Conv2d.weight[FLOAT, 32x8x1x1]\n  %module_list.18.BatchNorm2d.bias[FLOAT, 32]\n  %module_list.18.BatchNorm2d.running_mean[FLOAT, 32]\n  %module_list.18.BatchNorm2d.running_var[FLOAT, 32]\n  %module_list.18.BatchNorm2d.weight[FLOAT, 32]\n  %module_list.18.Conv2d.weight[FLOAT, 32x1x3x3]\n  %module_list.19.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.19.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.19.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.19.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.19.Conv2d.weight[FLOAT, 8x32x1x1]\n  %module_list.2.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.2.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.2.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.2.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.2.Conv2d.weight[FLOAT, 8x1x3x3]\n  %module_list.22.BatchNorm2d.bias[FLOAT, 32]\n  %module_list.22.BatchNorm2d.running_mean[FLOAT, 32]\n  %module_list.22.BatchNorm2d.running_var[FLOAT, 32]\n  %module_list.22.BatchNorm2d.weight[FLOAT, 32]\n  %module_list.22.Conv2d.weight[FLOAT, 32x8x1x1]\n  %module_list.23.BatchNorm2d.bias[FLOAT, 32]\n  %module_list.23.BatchNorm2d.running_mean[FLOAT, 32]\n  %module_list.23.BatchNorm2d.running_var[FLOAT, 32]\n  %module_list.23.BatchNorm2d.weight[FLOAT, 32]\n  %module_list.23.Conv2d.weight[FLOAT, 32x1x3x3]\n  %module_list.24.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.24.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.24.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.24.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.24.Conv2d.weight[FLOAT, 8x32x1x1]\n  %module_list.25.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.25.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.25.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.25.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.25.Conv2d.weight[FLOAT, 48x8x1x1]\n  %module_list.26.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.26.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.26.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.26.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.26.Conv2d.weight[FLOAT, 48x1x3x3]\n  %module_list.27.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.27.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.27.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.27.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.27.Conv2d.weight[FLOAT, 8x48x1x1]\n  %module_list.3.BatchNorm2d.bias[FLOAT, 4]\n  %module_list.3.BatchNorm2d.running_mean[FLOAT, 4]\n  %module_list.3.BatchNorm2d.running_var[FLOAT, 4]\n  %module_list.3.BatchNorm2d.weight[FLOAT, 4]\n  %module_list.3.Conv2d.weight[FLOAT, 4x8x1x1]\n  %module_list.30.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.30.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.30.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.30.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.30.Conv2d.weight[FLOAT, 48x8x1x1]\n  %module_list.31.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.31.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.31.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.31.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.31.Conv2d.weight[FLOAT, 48x1x3x3]\n  %module_list.32.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.32.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.32.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.32.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.32.Conv2d.weight[FLOAT, 8x48x1x1]\n  %module_list.35.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.35.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.35.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.35.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.35.Conv2d.weight[FLOAT, 48x8x1x1]\n  %module_list.36.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.36.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.36.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.36.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.36.Conv2d.weight[FLOAT, 48x1x3x3]\n  %module_list.37.BatchNorm2d.bias[FLOAT, 16]\n  %module_list.37.BatchNorm2d.running_mean[FLOAT, 16]\n  %module_list.37.BatchNorm2d.running_var[FLOAT, 16]\n  %module_list.37.BatchNorm2d.weight[FLOAT, 16]\n  %module_list.37.Conv2d.weight[FLOAT, 16x48x1x1]\n  %module_list.38.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.38.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.38.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.38.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.38.Conv2d.weight[FLOAT, 96x16x1x1]\n  %module_list.39.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.39.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.39.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.39.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.39.Conv2d.weight[FLOAT, 96x1x3x3]\n  %module_list.4.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.4.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.4.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.4.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.4.Conv2d.weight[FLOAT, 8x4x1x1]\n  %module_list.40.BatchNorm2d.bias[FLOAT, 16]\n  %module_list.40.BatchNorm2d.running_mean[FLOAT, 16]\n  %module_list.40.BatchNorm2d.running_var[FLOAT, 16]\n  %module_list.40.BatchNorm2d.weight[FLOAT, 16]\n  %module_list.40.Conv2d.weight[FLOAT, 16x96x1x1]\n  %module_list.43.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.43.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.43.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.43.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.43.Conv2d.weight[FLOAT, 96x16x1x1]\n  %module_list.44.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.44.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.44.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.44.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.44.Conv2d.weight[FLOAT, 96x1x3x3]\n  %module_list.45.BatchNorm2d.bias[FLOAT, 16]\n  %module_list.45.BatchNorm2d.running_mean[FLOAT, 16]\n  %module_list.45.BatchNorm2d.running_var[FLOAT, 16]\n  %module_list.45.BatchNorm2d.weight[FLOAT, 16]\n  %module_list.45.Conv2d.weight[FLOAT, 16x96x1x1]\n  %module_list.48.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.48.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.48.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.48.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.48.Conv2d.weight[FLOAT, 96x16x1x1]\n  %module_list.49.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.49.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.49.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.49.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.49.Conv2d.weight[FLOAT, 96x1x3x3]\n  %module_list.5.BatchNorm2d.bias[FLOAT, 8]\n  %module_list.5.BatchNorm2d.running_mean[FLOAT, 8]\n  %module_list.5.BatchNorm2d.running_var[FLOAT, 8]\n  %module_list.5.BatchNorm2d.weight[FLOAT, 8]\n  %module_list.5.Conv2d.weight[FLOAT, 8x1x3x3]\n  %module_list.50.BatchNorm2d.bias[FLOAT, 16]\n  %module_list.50.BatchNorm2d.running_mean[FLOAT, 16]\n  %module_list.50.BatchNorm2d.running_var[FLOAT, 16]\n  %module_list.50.BatchNorm2d.weight[FLOAT, 16]\n  %module_list.50.Conv2d.weight[FLOAT, 16x96x1x1]\n  %module_list.53.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.53.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.53.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.53.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.53.Conv2d.weight[FLOAT, 96x16x1x1]\n  %module_list.54.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.54.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.54.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.54.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.54.Conv2d.weight[FLOAT, 96x1x3x3]\n  %module_list.55.BatchNorm2d.bias[FLOAT, 16]\n  %module_list.55.BatchNorm2d.running_mean[FLOAT, 16]\n  %module_list.55.BatchNorm2d.running_var[FLOAT, 16]\n  %module_list.55.BatchNorm2d.weight[FLOAT, 16]\n  %module_list.55.Conv2d.weight[FLOAT, 16x96x1x1]\n  %module_list.58.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.58.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.58.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.58.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.58.Conv2d.weight[FLOAT, 96x16x1x1]\n  %module_list.59.BatchNorm2d.bias[FLOAT, 96]\n  %module_list.59.BatchNorm2d.running_mean[FLOAT, 96]\n  %module_list.59.BatchNorm2d.running_var[FLOAT, 96]\n  %module_list.59.BatchNorm2d.weight[FLOAT, 96]\n  %module_list.59.Conv2d.weight[FLOAT, 96x1x3x3]\n  %module_list.6.BatchNorm2d.bias[FLOAT, 4]\n  %module_list.6.BatchNorm2d.running_mean[FLOAT, 4]\n  %module_list.6.BatchNorm2d.running_var[FLOAT, 4]\n  %module_list.6.BatchNorm2d.weight[FLOAT, 4]\n  %module_list.6.Conv2d.weight[FLOAT, 4x8x1x1]\n  %module_list.60.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.60.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.60.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.60.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.60.Conv2d.weight[FLOAT, 24x96x1x1]\n  %module_list.61.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.61.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.61.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.61.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.61.Conv2d.weight[FLOAT, 136x24x1x1]\n  %module_list.62.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.62.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.62.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.62.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.62.Conv2d.weight[FLOAT, 136x1x3x3]\n  %module_list.63.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.63.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.63.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.63.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.63.Conv2d.weight[FLOAT, 24x136x1x1]\n  %module_list.66.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.66.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.66.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.66.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.66.Conv2d.weight[FLOAT, 136x24x1x1]\n  %module_list.67.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.67.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.67.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.67.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.67.Conv2d.weight[FLOAT, 136x1x3x3]\n  %module_list.68.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.68.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.68.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.68.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.68.Conv2d.weight[FLOAT, 24x136x1x1]\n  %module_list.71.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.71.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.71.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.71.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.71.Conv2d.weight[FLOAT, 136x24x1x1]\n  %module_list.72.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.72.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.72.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.72.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.72.Conv2d.weight[FLOAT, 136x1x3x3]\n  %module_list.73.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.73.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.73.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.73.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.73.Conv2d.weight[FLOAT, 24x136x1x1]\n  %module_list.76.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.76.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.76.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.76.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.76.Conv2d.weight[FLOAT, 136x24x1x1]\n  %module_list.77.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.77.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.77.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.77.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.77.Conv2d.weight[FLOAT, 136x1x3x3]\n  %module_list.78.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.78.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.78.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.78.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.78.Conv2d.weight[FLOAT, 24x136x1x1]\n  %module_list.81.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.81.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.81.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.81.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.81.Conv2d.weight[FLOAT, 136x24x1x1]\n  %module_list.82.BatchNorm2d.bias[FLOAT, 136]\n  %module_list.82.BatchNorm2d.running_mean[FLOAT, 136]\n  %module_list.82.BatchNorm2d.running_var[FLOAT, 136]\n  %module_list.82.BatchNorm2d.weight[FLOAT, 136]\n  %module_list.82.Conv2d.weight[FLOAT, 136x1x3x3]\n  %module_list.83.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.83.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.83.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.83.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.83.Conv2d.weight[FLOAT, 48x136x1x1]\n  %module_list.84.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.84.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.84.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.84.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.84.Conv2d.weight[FLOAT, 224x48x1x1]\n  %module_list.85.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.85.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.85.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.85.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.85.Conv2d.weight[FLOAT, 224x1x3x3]\n  %module_list.86.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.86.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.86.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.86.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.86.Conv2d.weight[FLOAT, 48x224x1x1]\n  %module_list.89.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.89.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.89.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.89.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.89.Conv2d.weight[FLOAT, 224x48x1x1]\n  %module_list.9.BatchNorm2d.bias[FLOAT, 24]\n  %module_list.9.BatchNorm2d.running_mean[FLOAT, 24]\n  %module_list.9.BatchNorm2d.running_var[FLOAT, 24]\n  %module_list.9.BatchNorm2d.weight[FLOAT, 24]\n  %module_list.9.Conv2d.weight[FLOAT, 24x4x1x1]\n  %module_list.90.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.90.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.90.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.90.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.90.Conv2d.weight[FLOAT, 224x1x3x3]\n  %module_list.91.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.91.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.91.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.91.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.91.Conv2d.weight[FLOAT, 48x224x1x1]\n  %module_list.94.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.94.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.94.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.94.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.94.Conv2d.weight[FLOAT, 224x48x1x1]\n  %module_list.95.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.95.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.95.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.95.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.95.Conv2d.weight[FLOAT, 224x1x3x3]\n  %module_list.96.BatchNorm2d.bias[FLOAT, 48]\n  %module_list.96.BatchNorm2d.running_mean[FLOAT, 48]\n  %module_list.96.BatchNorm2d.running_var[FLOAT, 48]\n  %module_list.96.BatchNorm2d.weight[FLOAT, 48]\n  %module_list.96.Conv2d.weight[FLOAT, 48x224x1x1]\n  %module_list.99.BatchNorm2d.bias[FLOAT, 224]\n  %module_list.99.BatchNorm2d.running_mean[FLOAT, 224]\n  %module_list.99.BatchNorm2d.running_var[FLOAT, 224]\n  %module_list.99.BatchNorm2d.weight[FLOAT, 224]\n  %module_list.99.Conv2d.weight[FLOAT, 224x48x1x1]\n) {\n  %503 = Conv[dilations = [1, 1], group = 1, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%images, %module_list.0.Conv2d.weight)\n  %504 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%503, %module_list.0.BatchNorm2d.weight, %module_list.0.BatchNorm2d.bias, %module_list.0.BatchNorm2d.running_mean, %module_list.0.BatchNorm2d.running_var)\n  %505 = LeakyRelu[alpha = 0.100000001490116](%504)\n  %506 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%505, %module_list.1.Conv2d.weight)\n  %507 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%506, %module_list.1.BatchNorm2d.weight, %module_list.1.BatchNorm2d.bias, %module_list.1.BatchNorm2d.running_mean, %module_list.1.BatchNorm2d.running_var)\n  %508 = LeakyRelu[alpha = 0.100000001490116](%507)\n  %509 = Conv[dilations = [1, 1], group = 8, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%508, %module_list.2.Conv2d.weight)\n  %510 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%509, %module_list.2.BatchNorm2d.weight, %module_list.2.BatchNorm2d.bias, %module_list.2.BatchNorm2d.running_mean, %module_list.2.BatchNorm2d.running_var)\n  %511 = LeakyRelu[alpha = 0.100000001490116](%510)\n  %512 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%511, %module_list.3.Conv2d.weight)\n  %513 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%512, %module_list.3.BatchNorm2d.weight, %module_list.3.BatchNorm2d.bias, %module_list.3.BatchNorm2d.running_mean, %module_list.3.BatchNorm2d.running_var)\n  %514 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%513, %module_list.4.Conv2d.weight)\n  %515 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%514, %module_list.4.BatchNorm2d.weight, %module_list.4.BatchNorm2d.bias, %module_list.4.BatchNorm2d.running_mean, %module_list.4.BatchNorm2d.running_var)\n  %516 = LeakyRelu[alpha = 0.100000001490116](%515)\n  %517 = Conv[dilations = [1, 1], group = 8, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%516, %module_list.5.Conv2d.weight)\n  %518 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%517, %module_list.5.BatchNorm2d.weight, %module_list.5.BatchNorm2d.bias, %module_list.5.BatchNorm2d.running_mean, %module_list.5.BatchNorm2d.running_var)\n  %519 = LeakyRelu[alpha = 0.100000001490116](%518)\n  %520 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%519, %module_list.6.Conv2d.weight)\n  %521 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%520, %module_list.6.BatchNorm2d.weight, %module_list.6.BatchNorm2d.bias, %module_list.6.BatchNorm2d.running_mean, %module_list.6.BatchNorm2d.running_var)\n  %522 = Add(%521, %513)\n  %523 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%522, %module_list.9.Conv2d.weight)\n  %524 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%523, %module_list.9.BatchNorm2d.weight, %module_list.9.BatchNorm2d.bias, %module_list.9.BatchNorm2d.running_mean, %module_list.9.BatchNorm2d.running_var)\n  %525 = LeakyRelu[alpha = 0.100000001490116](%524)\n  %526 = Conv[dilations = [1, 1], group = 24, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%525, %module_list.10.Conv2d.weight)\n  %527 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%526, %module_list.10.BatchNorm2d.weight, %module_list.10.BatchNorm2d.bias, %module_list.10.BatchNorm2d.running_mean, %module_list.10.BatchNorm2d.running_var)\n  %528 = LeakyRelu[alpha = 0.100000001490116](%527)\n  %529 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%528, %module_list.11.Conv2d.weight)\n  %530 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%529, %module_list.11.BatchNorm2d.weight, %module_list.11.BatchNorm2d.bias, %module_list.11.BatchNorm2d.running_mean, %module_list.11.BatchNorm2d.running_var)\n  %531 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%530, %module_list.12.Conv2d.weight)\n  %532 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%531, %module_list.12.BatchNorm2d.weight, %module_list.12.BatchNorm2d.bias, %module_list.12.BatchNorm2d.running_mean, %module_list.12.BatchNorm2d.running_var)\n  %533 = LeakyRelu[alpha = 0.100000001490116](%532)\n  %534 = Conv[dilations = [1, 1], group = 32, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%533, %module_list.13.Conv2d.weight)\n  %535 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%534, %module_list.13.BatchNorm2d.weight, %module_list.13.BatchNorm2d.bias, %module_list.13.BatchNorm2d.running_mean, %module_list.13.BatchNorm2d.running_var)\n  %536 = LeakyRelu[alpha = 0.100000001490116](%535)\n  %537 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%536, %module_list.14.Conv2d.weight)\n  %538 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%537, %module_list.14.BatchNorm2d.weight, %module_list.14.BatchNorm2d.bias, %module_list.14.BatchNorm2d.running_mean, %module_list.14.BatchNorm2d.running_var)\n  %539 = Add(%538, %530)\n  %540 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%539, %module_list.17.Conv2d.weight)\n  %541 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%540, %module_list.17.BatchNorm2d.weight, %module_list.17.BatchNorm2d.bias, %module_list.17.BatchNorm2d.running_mean, %module_list.17.BatchNorm2d.running_var)\n  %542 = LeakyRelu[alpha = 0.100000001490116](%541)\n  %543 = Conv[dilations = [1, 1], group = 32, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%542, %module_list.18.Conv2d.weight)\n  %544 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%543, %module_list.18.BatchNorm2d.weight, %module_list.18.BatchNorm2d.bias, %module_list.18.BatchNorm2d.running_mean, %module_list.18.BatchNorm2d.running_var)\n  %545 = LeakyRelu[alpha = 0.100000001490116](%544)\n  %546 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%545, %module_list.19.Conv2d.weight)\n  %547 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%546, %module_list.19.BatchNorm2d.weight, %module_list.19.BatchNorm2d.bias, %module_list.19.BatchNorm2d.running_mean, %module_list.19.BatchNorm2d.running_var)\n  %548 = Add(%547, %539)\n  %549 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%548, %module_list.22.Conv2d.weight)\n  %550 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%549, %module_list.22.BatchNorm2d.weight, %module_list.22.BatchNorm2d.bias, %module_list.22.BatchNorm2d.running_mean, %module_list.22.BatchNorm2d.running_var)\n  %551 = LeakyRelu[alpha = 0.100000001490116](%550)\n  %552 = Conv[dilations = [1, 1], group = 32, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%551, %module_list.23.Conv2d.weight)\n  %553 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%552, %module_list.23.BatchNorm2d.weight, %module_list.23.BatchNorm2d.bias, %module_list.23.BatchNorm2d.running_mean, %module_list.23.BatchNorm2d.running_var)\n  %554 = LeakyRelu[alpha = 0.100000001490116](%553)\n  %555 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%554, %module_list.24.Conv2d.weight)\n  %556 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%555, %module_list.24.BatchNorm2d.weight, %module_list.24.BatchNorm2d.bias, %module_list.24.BatchNorm2d.running_mean, %module_list.24.BatchNorm2d.running_var)\n  %557 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%556, %module_list.25.Conv2d.weight)\n  %558 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%557, %module_list.25.BatchNorm2d.weight, %module_list.25.BatchNorm2d.bias, %module_list.25.BatchNorm2d.running_mean, %module_list.25.BatchNorm2d.running_var)\n  %559 = LeakyRelu[alpha = 0.100000001490116](%558)\n  %560 = Conv[dilations = [1, 1], group = 48, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%559, %module_list.26.Conv2d.weight)\n  %561 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%560, %module_list.26.BatchNorm2d.weight, %module_list.26.BatchNorm2d.bias, %module_list.26.BatchNorm2d.running_mean, %module_list.26.BatchNorm2d.running_var)\n  %562 = LeakyRelu[alpha = 0.100000001490116](%561)\n  %563 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%562, %module_list.27.Conv2d.weight)\n  %564 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%563, %module_list.27.BatchNorm2d.weight, %module_list.27.BatchNorm2d.bias, %module_list.27.BatchNorm2d.running_mean, %module_list.27.BatchNorm2d.running_var)\n  %565 = Add(%564, %556)\n  %566 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%565, %module_list.30.Conv2d.weight)\n  %567 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%566, %module_list.30.BatchNorm2d.weight, %module_list.30.BatchNorm2d.bias, %module_list.30.BatchNorm2d.running_mean, %module_list.30.BatchNorm2d.running_var)\n  %568 = LeakyRelu[alpha = 0.100000001490116](%567)\n  %569 = Conv[dilations = [1, 1], group = 48, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%568, %module_list.31.Conv2d.weight)\n  %570 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%569, %module_list.31.BatchNorm2d.weight, %module_list.31.BatchNorm2d.bias, %module_list.31.BatchNorm2d.running_mean, %module_list.31.BatchNorm2d.running_var)\n  %571 = LeakyRelu[alpha = 0.100000001490116](%570)\n  %572 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%571, %module_list.32.Conv2d.weight)\n  %573 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%572, %module_list.32.BatchNorm2d.weight, %module_list.32.BatchNorm2d.bias, %module_list.32.BatchNorm2d.running_mean, %module_list.32.BatchNorm2d.running_var)\n  %574 = Add(%573, %565)\n  %575 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%574, %module_list.35.Conv2d.weight)\n  %576 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%575, %module_list.35.BatchNorm2d.weight, %module_list.35.BatchNorm2d.bias, %module_list.35.BatchNorm2d.running_mean, %module_list.35.BatchNorm2d.running_var)\n  %577 = LeakyRelu[alpha = 0.100000001490116](%576)\n  %578 = Conv[dilations = [1, 1], group = 48, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%577, %module_list.36.Conv2d.weight)\n  %579 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%578, %module_list.36.BatchNorm2d.weight, %module_list.36.BatchNorm2d.bias, %module_list.36.BatchNorm2d.running_mean, %module_list.36.BatchNorm2d.running_var)\n  %580 = LeakyRelu[alpha = 0.100000001490116](%579)\n  %581 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%580, %module_list.37.Conv2d.weight)\n  %582 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%581, %module_list.37.BatchNorm2d.weight, %module_list.37.BatchNorm2d.bias, %module_list.37.BatchNorm2d.running_mean, %module_list.37.BatchNorm2d.running_var)\n  %583 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%582, %module_list.38.Conv2d.weight)\n  %584 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%583, %module_list.38.BatchNorm2d.weight, %module_list.38.BatchNorm2d.bias, %module_list.38.BatchNorm2d.running_mean, %module_list.38.BatchNorm2d.running_var)\n  %585 = LeakyRelu[alpha = 0.100000001490116](%584)\n  %586 = Conv[dilations = [1, 1], group = 96, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%585, %module_list.39.Conv2d.weight)\n  %587 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%586, %module_list.39.BatchNorm2d.weight, %module_list.39.BatchNorm2d.bias, %module_list.39.BatchNorm2d.running_mean, %module_list.39.BatchNorm2d.running_var)\n  %588 = LeakyRelu[alpha = 0.100000001490116](%587)\n  %589 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%588, %module_list.40.Conv2d.weight)\n  %590 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%589, %module_list.40.BatchNorm2d.weight, %module_list.40.BatchNorm2d.bias, %module_list.40.BatchNorm2d.running_mean, %module_list.40.BatchNorm2d.running_var)\n  %591 = Add(%590, %582)\n  %592 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%591, %module_list.43.Conv2d.weight)\n  %593 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%592, %module_list.43.BatchNorm2d.weight, %module_list.43.BatchNorm2d.bias, %module_list.43.BatchNorm2d.running_mean, %module_list.43.BatchNorm2d.running_var)\n  %594 = LeakyRelu[alpha = 0.100000001490116](%593)\n  %595 = Conv[dilations = [1, 1], group = 96, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%594, %module_list.44.Conv2d.weight)\n  %596 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%595, %module_list.44.BatchNorm2d.weight, %module_list.44.BatchNorm2d.bias, %module_list.44.BatchNorm2d.running_mean, %module_list.44.BatchNorm2d.running_var)\n  %597 = LeakyRelu[alpha = 0.100000001490116](%596)\n  %598 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%597, %module_list.45.Conv2d.weight)\n  %599 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%598, %module_list.45.BatchNorm2d.weight, %module_list.45.BatchNorm2d.bias, %module_list.45.BatchNorm2d.running_mean, %module_list.45.BatchNorm2d.running_var)\n  %600 = Add(%599, %591)\n  %601 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%600, %module_list.48.Conv2d.weight)\n  %602 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%601, %module_list.48.BatchNorm2d.weight, %module_list.48.BatchNorm2d.bias, %module_list.48.BatchNorm2d.running_mean, %module_list.48.BatchNorm2d.running_var)\n  %603 = LeakyRelu[alpha = 0.100000001490116](%602)\n  %604 = Conv[dilations = [1, 1], group = 96, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%603, %module_list.49.Conv2d.weight)\n  %605 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%604, %module_list.49.BatchNorm2d.weight, %module_list.49.BatchNorm2d.bias, %module_list.49.BatchNorm2d.running_mean, %module_list.49.BatchNorm2d.running_var)\n  %606 = LeakyRelu[alpha = 0.100000001490116](%605)\n  %607 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%606, %module_list.50.Conv2d.weight)\n  %608 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%607, %module_list.50.BatchNorm2d.weight, %module_list.50.BatchNorm2d.bias, %module_list.50.BatchNorm2d.running_mean, %module_list.50.BatchNorm2d.running_var)\n  %609 = Add(%608, %600)\n  %610 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%609, %module_list.53.Conv2d.weight)\n  %611 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%610, %module_list.53.BatchNorm2d.weight, %module_list.53.BatchNorm2d.bias, %module_list.53.BatchNorm2d.running_mean, %module_list.53.BatchNorm2d.running_var)\n  %612 = LeakyRelu[alpha = 0.100000001490116](%611)\n  %613 = Conv[dilations = [1, 1], group = 96, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%612, %module_list.54.Conv2d.weight)\n  %614 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%613, %module_list.54.BatchNorm2d.weight, %module_list.54.BatchNorm2d.bias, %module_list.54.BatchNorm2d.running_mean, %module_list.54.BatchNorm2d.running_var)\n  %615 = LeakyRelu[alpha = 0.100000001490116](%614)\n  %616 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%615, %module_list.55.Conv2d.weight)\n  %617 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%616, %module_list.55.BatchNorm2d.weight, %module_list.55.BatchNorm2d.bias, %module_list.55.BatchNorm2d.running_mean, %module_list.55.BatchNorm2d.running_var)\n  %618 = Add(%617, %609)\n  %619 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%618, %module_list.58.Conv2d.weight)\n  %620 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%619, %module_list.58.BatchNorm2d.weight, %module_list.58.BatchNorm2d.bias, %module_list.58.BatchNorm2d.running_mean, %module_list.58.BatchNorm2d.running_var)\n  %621 = LeakyRelu[alpha = 0.100000001490116](%620)\n  %622 = Conv[dilations = [1, 1], group = 96, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%621, %module_list.59.Conv2d.weight)\n  %623 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%622, %module_list.59.BatchNorm2d.weight, %module_list.59.BatchNorm2d.bias, %module_list.59.BatchNorm2d.running_mean, %module_list.59.BatchNorm2d.running_var)\n  %624 = LeakyRelu[alpha = 0.100000001490116](%623)\n  %625 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%624, %module_list.60.Conv2d.weight)\n  %626 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%625, %module_list.60.BatchNorm2d.weight, %module_list.60.BatchNorm2d.bias, %module_list.60.BatchNorm2d.running_mean, %module_list.60.BatchNorm2d.running_var)\n  %627 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%626, %module_list.61.Conv2d.weight)\n  %628 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%627, %module_list.61.BatchNorm2d.weight, %module_list.61.BatchNorm2d.bias, %module_list.61.BatchNorm2d.running_mean, %module_list.61.BatchNorm2d.running_var)\n  %629 = LeakyRelu[alpha = 0.100000001490116](%628)\n  %630 = Conv[dilations = [1, 1], group = 136, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%629, %module_list.62.Conv2d.weight)\n  %631 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%630, %module_list.62.BatchNorm2d.weight, %module_list.62.BatchNorm2d.bias, %module_list.62.BatchNorm2d.running_mean, %module_list.62.BatchNorm2d.running_var)\n  %632 = LeakyRelu[alpha = 0.100000001490116](%631)\n  %633 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%632, %module_list.63.Conv2d.weight)\n  %634 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%633, %module_list.63.BatchNorm2d.weight, %module_list.63.BatchNorm2d.bias, %module_list.63.BatchNorm2d.running_mean, %module_list.63.BatchNorm2d.running_var)\n  %635 = Add(%634, %626)\n  %636 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%635, %module_list.66.Conv2d.weight)\n  %637 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%636, %module_list.66.BatchNorm2d.weight, %module_list.66.BatchNorm2d.bias, %module_list.66.BatchNorm2d.running_mean, %module_list.66.BatchNorm2d.running_var)\n  %638 = LeakyRelu[alpha = 0.100000001490116](%637)\n  %639 = Conv[dilations = [1, 1], group = 136, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%638, %module_list.67.Conv2d.weight)\n  %640 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%639, %module_list.67.BatchNorm2d.weight, %module_list.67.BatchNorm2d.bias, %module_list.67.BatchNorm2d.running_mean, %module_list.67.BatchNorm2d.running_var)\n  %641 = LeakyRelu[alpha = 0.100000001490116](%640)\n  %642 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%641, %module_list.68.Conv2d.weight)\n  %643 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%642, %module_list.68.BatchNorm2d.weight, %module_list.68.BatchNorm2d.bias, %module_list.68.BatchNorm2d.running_mean, %module_list.68.BatchNorm2d.running_var)\n  %644 = Add(%643, %635)\n  %645 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%644, %module_list.71.Conv2d.weight)\n  %646 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%645, %module_list.71.BatchNorm2d.weight, %module_list.71.BatchNorm2d.bias, %module_list.71.BatchNorm2d.running_mean, %module_list.71.BatchNorm2d.running_var)\n  %647 = LeakyRelu[alpha = 0.100000001490116](%646)\n  %648 = Conv[dilations = [1, 1], group = 136, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%647, %module_list.72.Conv2d.weight)\n  %649 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%648, %module_list.72.BatchNorm2d.weight, %module_list.72.BatchNorm2d.bias, %module_list.72.BatchNorm2d.running_mean, %module_list.72.BatchNorm2d.running_var)\n  %650 = LeakyRelu[alpha = 0.100000001490116](%649)\n  %651 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%650, %module_list.73.Conv2d.weight)\n  %652 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%651, %module_list.73.BatchNorm2d.weight, %module_list.73.BatchNorm2d.bias, %module_list.73.BatchNorm2d.running_mean, %module_list.73.BatchNorm2d.running_var)\n  %653 = Add(%652, %644)\n  %654 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%653, %module_list.76.Conv2d.weight)\n  %655 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%654, %module_list.76.BatchNorm2d.weight, %module_list.76.BatchNorm2d.bias, %module_list.76.BatchNorm2d.running_mean, %module_list.76.BatchNorm2d.running_var)\n  %656 = LeakyRelu[alpha = 0.100000001490116](%655)\n  %657 = Conv[dilations = [1, 1], group = 136, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%656, %module_list.77.Conv2d.weight)\n  %658 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%657, %module_list.77.BatchNorm2d.weight, %module_list.77.BatchNorm2d.bias, %module_list.77.BatchNorm2d.running_mean, %module_list.77.BatchNorm2d.running_var)\n  %659 = LeakyRelu[alpha = 0.100000001490116](%658)\n  %660 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%659, %module_list.78.Conv2d.weight)\n  %661 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%660, %module_list.78.BatchNorm2d.weight, %module_list.78.BatchNorm2d.bias, %module_list.78.BatchNorm2d.running_mean, %module_list.78.BatchNorm2d.running_var)\n  %662 = Add(%661, %653)\n  %663 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%662, %module_list.81.Conv2d.weight)\n  %664 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%663, %module_list.81.BatchNorm2d.weight, %module_list.81.BatchNorm2d.bias, %module_list.81.BatchNorm2d.running_mean, %module_list.81.BatchNorm2d.running_var)\n  %665 = LeakyRelu[alpha = 0.100000001490116](%664)\n  %666 = Conv[dilations = [1, 1], group = 136, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [2, 2]](%665, %module_list.82.Conv2d.weight)\n  %667 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%666, %module_list.82.BatchNorm2d.weight, %module_list.82.BatchNorm2d.bias, %module_list.82.BatchNorm2d.running_mean, %module_list.82.BatchNorm2d.running_var)\n  %668 = LeakyRelu[alpha = 0.100000001490116](%667)\n  %669 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%668, %module_list.83.Conv2d.weight)\n  %670 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%669, %module_list.83.BatchNorm2d.weight, %module_list.83.BatchNorm2d.bias, %module_list.83.BatchNorm2d.running_mean, %module_list.83.BatchNorm2d.running_var)\n  %671 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%670, %module_list.84.Conv2d.weight)\n  %672 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%671, %module_list.84.BatchNorm2d.weight, %module_list.84.BatchNorm2d.bias, %module_list.84.BatchNorm2d.running_mean, %module_list.84.BatchNorm2d.running_var)\n  %673 = LeakyRelu[alpha = 0.100000001490116](%672)\n  %674 = Conv[dilations = [1, 1], group = 224, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%673, %module_list.85.Conv2d.weight)\n  %675 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%674, %module_list.85.BatchNorm2d.weight, %module_list.85.BatchNorm2d.bias, %module_list.85.BatchNorm2d.running_mean, %module_list.85.BatchNorm2d.running_var)\n  %676 = LeakyRelu[alpha = 0.100000001490116](%675)\n  %677 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%676, %module_list.86.Conv2d.weight)\n  %678 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%677, %module_list.86.BatchNorm2d.weight, %module_list.86.BatchNorm2d.bias, %module_list.86.BatchNorm2d.running_mean, %module_list.86.BatchNorm2d.running_var)\n  %679 = Add(%678, %670)\n  %680 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%679, %module_list.89.Conv2d.weight)\n  %681 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%680, %module_list.89.BatchNorm2d.weight, %module_list.89.BatchNorm2d.bias, %module_list.89.BatchNorm2d.running_mean, %module_list.89.BatchNorm2d.running_var)\n  %682 = LeakyRelu[alpha = 0.100000001490116](%681)\n  %683 = Conv[dilations = [1, 1], group = 224, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%682, %module_list.90.Conv2d.weight)\n  %684 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%683, %module_list.90.BatchNorm2d.weight, %module_list.90.BatchNorm2d.bias, %module_list.90.BatchNorm2d.running_mean, %module_list.90.BatchNorm2d.running_var)\n  %685 = LeakyRelu[alpha = 0.100000001490116](%684)\n  %686 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%685, %module_list.91.Conv2d.weight)\n  %687 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%686, %module_list.91.BatchNorm2d.weight, %module_list.91.BatchNorm2d.bias, %module_list.91.BatchNorm2d.running_mean, %module_list.91.BatchNorm2d.running_var)\n  %688 = Add(%687, %679)\n  %689 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%688, %module_list.94.Conv2d.weight)\n  %690 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%689, %module_list.94.BatchNorm2d.weight, %module_list.94.BatchNorm2d.bias, %module_list.94.BatchNorm2d.running_mean, %module_list.94.BatchNorm2d.running_var)\n  %691 = LeakyRelu[alpha = 0.100000001490116](%690)\n  %692 = Conv[dilations = [1, 1], group = 224, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%691, %module_list.95.Conv2d.weight)\n  %693 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%692, %module_list.95.BatchNorm2d.weight, %module_list.95.BatchNorm2d.bias, %module_list.95.BatchNorm2d.running_mean, %module_list.95.BatchNorm2d.running_var)\n  %694 = LeakyRelu[alpha = 0.100000001490116](%693)\n  %695 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%694, %module_list.96.Conv2d.weight)\n  %696 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%695, %module_list.96.BatchNorm2d.weight, %module_list.96.BatchNorm2d.bias, %module_list.96.BatchNorm2d.running_mean, %module_list.96.BatchNorm2d.running_var)\n  %697 = Add(%696, %688)\n  %698 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%697, %module_list.99.Conv2d.weight)\n  %699 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%698, %module_list.99.BatchNorm2d.weight, %module_list.99.BatchNorm2d.bias, %module_list.99.BatchNorm2d.running_mean, %module_list.99.BatchNorm2d.running_var)\n  %700 = LeakyRelu[alpha = 0.100000001490116](%699)\n  %701 = Conv[dilations = [1, 1], group = 224, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%700, %module_list.100.Conv2d.weight)\n  %702 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%701, %module_list.100.BatchNorm2d.weight, %module_list.100.BatchNorm2d.bias, %module_list.100.BatchNorm2d.running_mean, %module_list.100.BatchNorm2d.running_var)\n  %703 = LeakyRelu[alpha = 0.100000001490116](%702)\n  %704 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%703, %module_list.101.Conv2d.weight)\n  %705 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%704, %module_list.101.BatchNorm2d.weight, %module_list.101.BatchNorm2d.bias, %module_list.101.BatchNorm2d.running_mean, %module_list.101.BatchNorm2d.running_var)\n  %706 = Add(%705, %697)\n  %707 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%706, %module_list.104.Conv2d.weight)\n  %708 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%707, %module_list.104.BatchNorm2d.weight, %module_list.104.BatchNorm2d.bias, %module_list.104.BatchNorm2d.running_mean, %module_list.104.BatchNorm2d.running_var)\n  %709 = LeakyRelu[alpha = 0.100000001490116](%708)\n  %710 = Conv[dilations = [1, 1], group = 224, kernel_shape = [3, 3], pads = [1, 1, 1, 1], strides = [1, 1]](%709, %module_list.105.Conv2d.weight)\n  %711 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%710, %module_list.105.BatchNorm2d.weight, %module_list.105.BatchNorm2d.bias, %module_list.105.BatchNorm2d.running_mean, %module_list.105.BatchNorm2d.running_var)\n  %712 = LeakyRelu[alpha = 0.100000001490116](%711)\n  %713 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%712, %module_list.106.Conv2d.weight)\n  %714 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%713, %module_list.106.BatchNorm2d.weight, %module_list.106.BatchNorm2d.bias, %module_list.106.BatchNorm2d.running_mean, %module_list.106.BatchNorm2d.running_var)\n  %715 = Add(%714, %706)\n  %716 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%715, %module_list.109.Conv2d.weight)\n  %717 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%716, %module_list.109.BatchNorm2d.weight, %module_list.109.BatchNorm2d.bias, %module_list.109.BatchNorm2d.running_mean, %module_list.109.BatchNorm2d.running_var)\n  %718 = LeakyRelu[alpha = 0.100000001490116](%717)\n  %719 = Conv[dilations = [1, 1], group = 96, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%718, %module_list.110.Conv2d.weight)\n  %720 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%719, %module_list.110.BatchNorm2d.weight, %module_list.110.BatchNorm2d.bias, %module_list.110.BatchNorm2d.running_mean, %module_list.110.BatchNorm2d.running_var)\n  %721 = LeakyRelu[alpha = 0.100000001490116](%720)\n  %722 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%721, %module_list.111.Conv2d.weight)\n  %723 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%722, %module_list.111.BatchNorm2d.weight, %module_list.111.BatchNorm2d.bias, %module_list.111.BatchNorm2d.running_mean, %module_list.111.BatchNorm2d.running_var)\n  %724 = Conv[dilations = [1, 1], group = 128, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%723, %module_list.112.Conv2d.weight)\n  %725 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%724, %module_list.112.BatchNorm2d.weight, %module_list.112.BatchNorm2d.bias, %module_list.112.BatchNorm2d.running_mean, %module_list.112.BatchNorm2d.running_var)\n  %726 = LeakyRelu[alpha = 0.100000001490116](%725)\n  %727 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%726, %module_list.113.Conv2d.weight)\n  %728 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%727, %module_list.113.BatchNorm2d.weight, %module_list.113.BatchNorm2d.bias, %module_list.113.BatchNorm2d.running_mean, %module_list.113.BatchNorm2d.running_var)\n  %729 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%728, %module_list.114.Conv2d.weight, %module_list.114.Conv2d.bias)\n  %730 = Shape(%729)\n  %731 = Constant[value = <Scalar Tensor []>]()\n  %732 = Gather[axis = 0](%730, %731)\n  %737 = Unsqueeze[axes = [0]](%732)\n  %742 = Concat[axis = 0](%737, %1536, %1537, %1538, %1539)\n  %743 = Reshape(%729, %742)\n  %yolo1 = Transpose[perm = [0, 1, 3, 4, 2]](%743)\n  %745 = Constant[value = <Tensor>]()\n  %746 = Constant[value = <Tensor>]()\n  %747 = Constant[value = <Tensor>]()\n  %748 = Constant[value = <Tensor>]()\n  %749 = Slice(%yolo1, %746, %747, %745, %748)\n  %750 = Sigmoid(%749)\n  %751 = Constant[value = <Tensor>]()\n  %752 = Add(%750, %751)\n  %753 = Constant[value = <Tensor>]()\n  %754 = Reshape(%752, %753)\n  %755 = Constant[value = <Tensor>]()\n  %757 = ConstantOfShape[value = <Tensor>](%1540)\n  %758 = Constant[value = <Scalar Tensor []>]()\n  %759 = Mul(%757, %758)\n  %760 = Constant[value = <Tensor>]()\n  %761 = Equal(%760, %759)\n  %762 = Where(%761, %757, %755)\n  %763 = Expand(%754, %762)\n  %764 = Shape(%yolo1)\n  %765 = Constant[value = <Scalar Tensor []>]()\n  %766 = Gather[axis = 0](%764, %765)\n  %767 = Cast[to = 7](%766)\n  %768 = Constant[value = <Scalar Tensor []>]()\n  %769 = Constant[value = <Scalar Tensor []>]()\n  %770 = Range(%768, %767, %769)\n  %771 = Shape(%yolo1)\n  %772 = Constant[value = <Scalar Tensor []>]()\n  %773 = Gather[axis = 0](%771, %772)\n  %774 = Cast[to = 7](%773)\n  %775 = Constant[value = <Scalar Tensor []>]()\n  %776 = Constant[value = <Scalar Tensor []>]()\n  %777 = Range(%775, %774, %776)\n  %778 = Shape(%yolo1)\n  %779 = Constant[value = <Scalar Tensor []>]()\n  %780 = Gather[axis = 0](%778, %779)\n  %781 = Cast[to = 7](%780)\n  %782 = Constant[value = <Scalar Tensor []>]()\n  %783 = Constant[value = <Scalar Tensor []>]()\n  %784 = Range(%782, %781, %783)\n  %785 = Shape(%yolo1)\n  %786 = Constant[value = <Scalar Tensor []>]()\n  %787 = Gather[axis = 0](%785, %786)\n  %788 = Cast[to = 7](%787)\n  %789 = Constant[value = <Scalar Tensor []>]()\n  %790 = Constant[value = <Scalar Tensor []>]()\n  %791 = Range(%789, %788, %790)\n  %792 = Shape(%yolo1)\n  %793 = Constant[value = <Scalar Tensor []>]()\n  %794 = Gather[axis = 0](%792, %793)\n  %795 = Cast[to = 7](%794)\n  %796 = Constant[value = <Scalar Tensor []>]()\n  %797 = Constant[value = <Scalar Tensor []>]()\n  %798 = Range(%796, %795, %797)\n  %799 = Constant[value = <Tensor>]()\n  %800 = Constant[value = <Tensor>]()\n  %801 = Constant[value = <Tensor>]()\n  %802 = Constant[value = <Tensor>]()\n  %803 = Slice(%798, %800, %801, %799, %802)\n  %804 = Constant[value = <Tensor>]()\n  %805 = Reshape(%770, %804)\n  %806 = Constant[value = <Tensor>]()\n  %807 = Reshape(%777, %806)\n  %808 = Constant[value = <Tensor>]()\n  %809 = Reshape(%784, %808)\n  %810 = Constant[value = <Tensor>]()\n  %811 = Reshape(%791, %810)\n  %812 = Constant[value = <Tensor>]()\n  %813 = Reshape(%803, %812)\n  %814 = Add(%805, %807)\n  %815 = Add(%814, %809)\n  %816 = Add(%815, %811)\n  %817 = Add(%816, %813)\n  %818 = Shape(%817)\n  %819 = Shape(%818)\n  %820 = ConstantOfShape[value = <Tensor>](%819)\n  %821 = Constant[value = <Scalar Tensor []>]()\n  %822 = Mul(%820, %821)\n  %823 = Equal(%818, %822)\n  %824 = Where(%823, %820, %818)\n  %825 = Expand(%805, %824)\n  %826 = Unsqueeze[axes = [-1]](%825)\n  %827 = Shape(%818)\n  %828 = ConstantOfShape[value = <Tensor>](%827)\n  %829 = Constant[value = <Scalar Tensor []>]()\n  %830 = Mul(%828, %829)\n  %831 = Equal(%818, %830)\n  %832 = Where(%831, %828, %818)\n  %833 = Expand(%807, %832)\n  %834 = Unsqueeze[axes = [-1]](%833)\n  %835 = Shape(%818)\n  %836 = ConstantOfShape[value = <Tensor>](%835)\n  %837 = Constant[value = <Scalar Tensor []>]()\n  %838 = Mul(%836, %837)\n  %839 = Equal(%818, %838)\n  %840 = Where(%839, %836, %818)\n  %841 = Expand(%809, %840)\n  %842 = Unsqueeze[axes = [-1]](%841)\n  %843 = Shape(%818)\n  %844 = ConstantOfShape[value = <Tensor>](%843)\n  %845 = Constant[value = <Scalar Tensor []>]()\n  %846 = Mul(%844, %845)\n  %847 = Equal(%818, %846)\n  %848 = Where(%847, %844, %818)\n  %849 = Expand(%811, %848)\n  %850 = Unsqueeze[axes = [-1]](%849)\n  %851 = Shape(%818)\n  %852 = ConstantOfShape[value = <Tensor>](%851)\n  %853 = Constant[value = <Scalar Tensor []>]()\n  %854 = Mul(%852, %853)\n  %855 = Equal(%818, %854)\n  %856 = Where(%855, %852, %818)\n  %857 = Expand(%813, %856)\n  %858 = Unsqueeze[axes = [-1]](%857)\n  %859 = Concat[axis = -1](%826, %834, %842, %850, %858)\n  %860 = Shape(%yolo1)\n  %861 = Constant[value = <Tensor>]()\n  %862 = Constant[value = <Tensor>]()\n  %863 = Constant[value = <Tensor>]()\n  %864 = Slice(%860, %862, %863, %861)\n  %865 = Concat[axis = 0](%818, %864)\n  %866 = Reshape(%763, %865)\n  %867 = ScatterND(%yolo1, %859, %866)\n  %868 = Constant[value = <Tensor>]()\n  %869 = Constant[value = <Tensor>]()\n  %870 = Constant[value = <Tensor>]()\n  %871 = Constant[value = <Tensor>]()\n  %872 = Slice(%867, %869, %870, %868, %871)\n  %873 = Exp(%872)\n  %874 = Constant[value = <Tensor>]()\n  %875 = Mul(%873, %874)\n  %876 = Constant[value = <Tensor>]()\n  %877 = Reshape(%875, %876)\n  %878 = Constant[value = <Tensor>]()\n  %880 = ConstantOfShape[value = <Tensor>](%1541)\n  %881 = Constant[value = <Scalar Tensor []>]()\n  %882 = Mul(%880, %881)\n  %883 = Constant[value = <Tensor>]()\n  %884 = Equal(%883, %882)\n  %885 = Where(%884, %880, %878)\n  %886 = Expand(%877, %885)\n  %887 = Shape(%867)\n  %888 = Constant[value = <Scalar Tensor []>]()\n  %889 = Gather[axis = 0](%887, %888)\n  %890 = Cast[to = 7](%889)\n  %891 = Constant[value = <Scalar Tensor []>]()\n  %892 = Constant[value = <Scalar Tensor []>]()\n  %893 = Range(%891, %890, %892)\n  %894 = Shape(%867)\n  %895 = Constant[value = <Scalar Tensor []>]()\n  %896 = Gather[axis = 0](%894, %895)\n  %897 = Cast[to = 7](%896)\n  %898 = Constant[value = <Scalar Tensor []>]()\n  %899 = Constant[value = <Scalar Tensor []>]()\n  %900 = Range(%898, %897, %899)\n  %901 = Shape(%867)\n  %902 = Constant[value = <Scalar Tensor []>]()\n  %903 = Gather[axis = 0](%901, %902)\n  %904 = Cast[to = 7](%903)\n  %905 = Constant[value = <Scalar Tensor []>]()\n  %906 = Constant[value = <Scalar Tensor []>]()\n  %907 = Range(%905, %904, %906)\n  %908 = Shape(%867)\n  %909 = Constant[value = <Scalar Tensor []>]()\n  %910 = Gather[axis = 0](%908, %909)\n  %911 = Cast[to = 7](%910)\n  %912 = Constant[value = <Scalar Tensor []>]()\n  %913 = Constant[value = <Scalar Tensor []>]()\n  %914 = Range(%912, %911, %913)\n  %915 = Shape(%867)\n  %916 = Constant[value = <Scalar Tensor []>]()\n  %917 = Gather[axis = 0](%915, %916)\n  %918 = Cast[to = 7](%917)\n  %919 = Constant[value = <Scalar Tensor []>]()\n  %920 = Constant[value = <Scalar Tensor []>]()\n  %921 = Range(%919, %918, %920)\n  %922 = Constant[value = <Tensor>]()\n  %923 = Constant[value = <Tensor>]()\n  %924 = Constant[value = <Tensor>]()\n  %925 = Constant[value = <Tensor>]()\n  %926 = Slice(%921, %923, %924, %922, %925)\n  %927 = Constant[value = <Tensor>]()\n  %928 = Reshape(%893, %927)\n  %929 = Constant[value = <Tensor>]()\n  %930 = Reshape(%900, %929)\n  %931 = Constant[value = <Tensor>]()\n  %932 = Reshape(%907, %931)\n  %933 = Constant[value = <Tensor>]()\n  %934 = Reshape(%914, %933)\n  %935 = Constant[value = <Tensor>]()\n  %936 = Reshape(%926, %935)\n  %937 = Add(%928, %930)\n  %938 = Add(%937, %932)\n  %939 = Add(%938, %934)\n  %940 = Add(%939, %936)\n  %941 = Shape(%940)\n  %942 = Shape(%941)\n  %943 = ConstantOfShape[value = <Tensor>](%942)\n  %944 = Constant[value = <Scalar Tensor []>]()\n  %945 = Mul(%943, %944)\n  %946 = Equal(%941, %945)\n  %947 = Where(%946, %943, %941)\n  %948 = Expand(%928, %947)\n  %949 = Unsqueeze[axes = [-1]](%948)\n  %950 = Shape(%941)\n  %951 = ConstantOfShape[value = <Tensor>](%950)\n  %952 = Constant[value = <Scalar Tensor []>]()\n  %953 = Mul(%951, %952)\n  %954 = Equal(%941, %953)\n  %955 = Where(%954, %951, %941)\n  %956 = Expand(%930, %955)\n  %957 = Unsqueeze[axes = [-1]](%956)\n  %958 = Shape(%941)\n  %959 = ConstantOfShape[value = <Tensor>](%958)\n  %960 = Constant[value = <Scalar Tensor []>]()\n  %961 = Mul(%959, %960)\n  %962 = Equal(%941, %961)\n  %963 = Where(%962, %959, %941)\n  %964 = Expand(%932, %963)\n  %965 = Unsqueeze[axes = [-1]](%964)\n  %966 = Shape(%941)\n  %967 = ConstantOfShape[value = <Tensor>](%966)\n  %968 = Constant[value = <Scalar Tensor []>]()\n  %969 = Mul(%967, %968)\n  %970 = Equal(%941, %969)\n  %971 = Where(%970, %967, %941)\n  %972 = Expand(%934, %971)\n  %973 = Unsqueeze[axes = [-1]](%972)\n  %974 = Shape(%941)\n  %975 = ConstantOfShape[value = <Tensor>](%974)\n  %976 = Constant[value = <Scalar Tensor []>]()\n  %977 = Mul(%975, %976)\n  %978 = Equal(%941, %977)\n  %979 = Where(%978, %975, %941)\n  %980 = Expand(%936, %979)\n  %981 = Unsqueeze[axes = [-1]](%980)\n  %982 = Concat[axis = -1](%949, %957, %965, %973, %981)\n  %983 = Shape(%867)\n  %984 = Constant[value = <Tensor>]()\n  %985 = Constant[value = <Tensor>]()\n  %986 = Constant[value = <Tensor>]()\n  %987 = Slice(%983, %985, %986, %984)\n  %988 = Concat[axis = 0](%941, %987)\n  %989 = Reshape(%886, %988)\n  %990 = ScatterND(%867, %982, %989)\n  %991 = Constant[value = <Tensor>]()\n  %992 = Constant[value = <Tensor>]()\n  %993 = Constant[value = <Tensor>]()\n  %994 = Constant[value = <Tensor>]()\n  %995 = Slice(%990, %992, %993, %991, %994)\n  %996 = Constant[value = <Scalar Tensor []>]()\n  %997 = Mul(%995, %996)\n  %998 = Constant[value = <Tensor>]()\n  %999 = Reshape(%997, %998)\n  %1000 = Constant[value = <Tensor>]()\n  %1002 = ConstantOfShape[value = <Tensor>](%1542)\n  %1003 = Constant[value = <Scalar Tensor []>]()\n  %1004 = Mul(%1002, %1003)\n  %1005 = Constant[value = <Tensor>]()\n  %1006 = Equal(%1005, %1004)\n  %1007 = Where(%1006, %1002, %1000)\n  %1008 = Expand(%999, %1007)\n  %1009 = Shape(%990)\n  %1010 = Constant[value = <Scalar Tensor []>]()\n  %1011 = Gather[axis = 0](%1009, %1010)\n  %1012 = Cast[to = 7](%1011)\n  %1013 = Constant[value = <Scalar Tensor []>]()\n  %1014 = Constant[value = <Scalar Tensor []>]()\n  %1015 = Range(%1013, %1012, %1014)\n  %1016 = Shape(%990)\n  %1017 = Constant[value = <Scalar Tensor []>]()\n  %1018 = Gather[axis = 0](%1016, %1017)\n  %1019 = Cast[to = 7](%1018)\n  %1020 = Constant[value = <Scalar Tensor []>]()\n  %1021 = Constant[value = <Scalar Tensor []>]()\n  %1022 = Range(%1020, %1019, %1021)\n  %1023 = Shape(%990)\n  %1024 = Constant[value = <Scalar Tensor []>]()\n  %1025 = Gather[axis = 0](%1023, %1024)\n  %1026 = Cast[to = 7](%1025)\n  %1027 = Constant[value = <Scalar Tensor []>]()\n  %1028 = Constant[value = <Scalar Tensor []>]()\n  %1029 = Range(%1027, %1026, %1028)\n  %1030 = Shape(%990)\n  %1031 = Constant[value = <Scalar Tensor []>]()\n  %1032 = Gather[axis = 0](%1030, %1031)\n  %1033 = Cast[to = 7](%1032)\n  %1034 = Constant[value = <Scalar Tensor []>]()\n  %1035 = Constant[value = <Scalar Tensor []>]()\n  %1036 = Range(%1034, %1033, %1035)\n  %1037 = Shape(%990)\n  %1038 = Constant[value = <Scalar Tensor []>]()\n  %1039 = Gather[axis = 0](%1037, %1038)\n  %1040 = Cast[to = 7](%1039)\n  %1041 = Constant[value = <Scalar Tensor []>]()\n  %1042 = Constant[value = <Scalar Tensor []>]()\n  %1043 = Range(%1041, %1040, %1042)\n  %1044 = Constant[value = <Tensor>]()\n  %1045 = Constant[value = <Tensor>]()\n  %1046 = Constant[value = <Tensor>]()\n  %1047 = Constant[value = <Tensor>]()\n  %1048 = Slice(%1043, %1045, %1046, %1044, %1047)\n  %1049 = Constant[value = <Tensor>]()\n  %1050 = Reshape(%1015, %1049)\n  %1051 = Constant[value = <Tensor>]()\n  %1052 = Reshape(%1022, %1051)\n  %1053 = Constant[value = <Tensor>]()\n  %1054 = Reshape(%1029, %1053)\n  %1055 = Constant[value = <Tensor>]()\n  %1056 = Reshape(%1036, %1055)\n  %1057 = Constant[value = <Tensor>]()\n  %1058 = Reshape(%1048, %1057)\n  %1059 = Add(%1050, %1052)\n  %1060 = Add(%1059, %1054)\n  %1061 = Add(%1060, %1056)\n  %1062 = Add(%1061, %1058)\n  %1063 = Shape(%1062)\n  %1064 = Shape(%1063)\n  %1065 = ConstantOfShape[value = <Tensor>](%1064)\n  %1066 = Constant[value = <Scalar Tensor []>]()\n  %1067 = Mul(%1065, %1066)\n  %1068 = Equal(%1063, %1067)\n  %1069 = Where(%1068, %1065, %1063)\n  %1070 = Expand(%1050, %1069)\n  %1071 = Unsqueeze[axes = [-1]](%1070)\n  %1072 = Shape(%1063)\n  %1073 = ConstantOfShape[value = <Tensor>](%1072)\n  %1074 = Constant[value = <Scalar Tensor []>]()\n  %1075 = Mul(%1073, %1074)\n  %1076 = Equal(%1063, %1075)\n  %1077 = Where(%1076, %1073, %1063)\n  %1078 = Expand(%1052, %1077)\n  %1079 = Unsqueeze[axes = [-1]](%1078)\n  %1080 = Shape(%1063)\n  %1081 = ConstantOfShape[value = <Tensor>](%1080)\n  %1082 = Constant[value = <Scalar Tensor []>]()\n  %1083 = Mul(%1081, %1082)\n  %1084 = Equal(%1063, %1083)\n  %1085 = Where(%1084, %1081, %1063)\n  %1086 = Expand(%1054, %1085)\n  %1087 = Unsqueeze[axes = [-1]](%1086)\n  %1088 = Shape(%1063)\n  %1089 = ConstantOfShape[value = <Tensor>](%1088)\n  %1090 = Constant[value = <Scalar Tensor []>]()\n  %1091 = Mul(%1089, %1090)\n  %1092 = Equal(%1063, %1091)\n  %1093 = Where(%1092, %1089, %1063)\n  %1094 = Expand(%1056, %1093)\n  %1095 = Unsqueeze[axes = [-1]](%1094)\n  %1096 = Shape(%1063)\n  %1097 = ConstantOfShape[value = <Tensor>](%1096)\n  %1098 = Constant[value = <Scalar Tensor []>]()\n  %1099 = Mul(%1097, %1098)\n  %1100 = Equal(%1063, %1099)\n  %1101 = Where(%1100, %1097, %1063)\n  %1102 = Expand(%1058, %1101)\n  %1103 = Unsqueeze[axes = [-1]](%1102)\n  %1104 = Concat[axis = -1](%1071, %1079, %1087, %1095, %1103)\n  %1105 = Shape(%990)\n  %1106 = Constant[value = <Tensor>]()\n  %1107 = Constant[value = <Tensor>]()\n  %1108 = Constant[value = <Tensor>]()\n  %1109 = Slice(%1105, %1107, %1108, %1106)\n  %1110 = Concat[axis = 0](%1063, %1109)\n  %1111 = Reshape(%1008, %1110)\n  %1112 = ScatterND(%990, %1104, %1111)\n  %1115 = Unsqueeze[axes = [0]](%732)\n  %1118 = Concat[axis = 0](%1115, %1543, %1544)\n  %1119 = Reshape(%1112, %1118)\n  %1128 = Constant[value = <Tensor>]()\n  %1129 = Resize[coordinate_transformation_mode = 'asymmetric', cubic_coeff_a = -0.75, mode = 'nearest', nearest_mode = 'floor'](%718, %1128, %1549)\n  %1130 = Concat[axis = 1](%1129, %665)\n  %1131 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1130, %module_list.119.Conv2d.weight)\n  %1132 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1131, %module_list.119.BatchNorm2d.weight, %module_list.119.BatchNorm2d.bias, %module_list.119.BatchNorm2d.running_mean, %module_list.119.BatchNorm2d.running_var)\n  %1133 = LeakyRelu[alpha = 0.100000001490116](%1132)\n  %1134 = Conv[dilations = [1, 1], group = 96, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%1133, %module_list.120.Conv2d.weight)\n  %1135 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1134, %module_list.120.BatchNorm2d.weight, %module_list.120.BatchNorm2d.bias, %module_list.120.BatchNorm2d.running_mean, %module_list.120.BatchNorm2d.running_var)\n  %1136 = LeakyRelu[alpha = 0.100000001490116](%1135)\n  %1137 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1136, %module_list.121.Conv2d.weight)\n  %1138 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1137, %module_list.121.BatchNorm2d.weight, %module_list.121.BatchNorm2d.bias, %module_list.121.BatchNorm2d.running_mean, %module_list.121.BatchNorm2d.running_var)\n  %1139 = Conv[dilations = [1, 1], group = 96, kernel_shape = [5, 5], pads = [2, 2, 2, 2], strides = [1, 1]](%1138, %module_list.122.Conv2d.weight)\n  %1140 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1139, %module_list.122.BatchNorm2d.weight, %module_list.122.BatchNorm2d.bias, %module_list.122.BatchNorm2d.running_mean, %module_list.122.BatchNorm2d.running_var)\n  %1141 = LeakyRelu[alpha = 0.100000001490116](%1140)\n  %1142 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1141, %module_list.123.Conv2d.weight)\n  %1143 = BatchNormalization[epsilon = 9.99999974737875e-05, momentum = 0.970000028610229](%1142, %module_list.123.BatchNorm2d.weight, %module_list.123.BatchNorm2d.bias, %module_list.123.BatchNorm2d.running_mean, %module_list.123.BatchNorm2d.running_var)\n  %1144 = Conv[dilations = [1, 1], group = 1, kernel_shape = [1, 1], pads = [0, 0, 0, 0], strides = [1, 1]](%1143, %module_list.124.Conv2d.weight, %module_list.124.Conv2d.bias)\n  %1145 = Shape(%1144)\n  %1146 = Constant[value = <Scalar Tensor []>]()\n  %1147 = Gather[axis = 0](%1145, %1146)\n  %1152 = Unsqueeze[axes = [0]](%1147)\n  %1157 = Concat[axis = 0](%1152, %1550, %1551, %1552, %1553)\n  %1158 = Reshape(%1144, %1157)\n  %yolo2 = Transpose[perm = [0, 1, 3, 4, 2]](%1158)\n  %1160 = Constant[value = <Tensor>]()\n  %1161 = Constant[value = <Tensor>]()\n  %1162 = Constant[value = <Tensor>]()\n  %1163 = Constant[value = <Tensor>]()\n  %1164 = Slice(%yolo2, %1161, %1162, %1160, %1163)\n  %1165 = Sigmoid(%1164)\n  %1166 = Constant[value = <Tensor>]()\n  %1167 = Add(%1165, %1166)\n  %1168 = Constant[value = <Tensor>]()\n  %1169 = Reshape(%1167, %1168)\n  %1170 = Constant[value = <Tensor>]()\n  %1172 = ConstantOfShape[value = <Tensor>](%1554)\n  %1173 = Constant[value = <Scalar Tensor []>]()\n  %1174 = Mul(%1172, %1173)\n  %1175 = Constant[value = <Tensor>]()\n  %1176 = Equal(%1175, %1174)\n  %1177 = Where(%1176, %1172, %1170)\n  %1178 = Expand(%1169, %1177)\n  %1179 = Shape(%yolo2)\n  %1180 = Constant[value = <Scalar Tensor []>]()\n  %1181 = Gather[axis = 0](%1179, %1180)\n  %1182 = Cast[to = 7](%1181)\n  %1183 = Constant[value = <Scalar Tensor []>]()\n  %1184 = Constant[value = <Scalar Tensor []>]()\n  %1185 = Range(%1183, %1182, %1184)\n  %1186 = Shape(%yolo2)\n  %1187 = Constant[value = <Scalar Tensor []>]()\n  %1188 = Gather[axis = 0](%1186, %1187)\n  %1189 = Cast[to = 7](%1188)\n  %1190 = Constant[value = <Scalar Tensor []>]()\n  %1191 = Constant[value = <Scalar Tensor []>]()\n  %1192 = Range(%1190, %1189, %1191)\n  %1193 = Shape(%yolo2)\n  %1194 = Constant[value = <Scalar Tensor []>]()\n  %1195 = Gather[axis = 0](%1193, %1194)\n  %1196 = Cast[to = 7](%1195)\n  %1197 = Constant[value = <Scalar Tensor []>]()\n  %1198 = Constant[value = <Scalar Tensor []>]()\n  %1199 = Range(%1197, %1196, %1198)\n  %1200 = Shape(%yolo2)\n  %1201 = Constant[value = <Scalar Tensor []>]()\n  %1202 = Gather[axis = 0](%1200, %1201)\n  %1203 = Cast[to = 7](%1202)\n  %1204 = Constant[value = <Scalar Tensor []>]()\n  %1205 = Constant[value = <Scalar Tensor []>]()\n  %1206 = Range(%1204, %1203, %1205)\n  %1207 = Shape(%yolo2)\n  %1208 = Constant[value = <Scalar Tensor []>]()\n  %1209 = Gather[axis = 0](%1207, %1208)\n  %1210 = Cast[to = 7](%1209)\n  %1211 = Constant[value = <Scalar Tensor []>]()\n  %1212 = Constant[value = <Scalar Tensor []>]()\n  %1213 = Range(%1211, %1210, %1212)\n  %1214 = Constant[value = <Tensor>]()\n  %1215 = Constant[value = <Tensor>]()\n  %1216 = Constant[value = <Tensor>]()\n  %1217 = Constant[value = <Tensor>]()\n  %1218 = Slice(%1213, %1215, %1216, %1214, %1217)\n  %1219 = Constant[value = <Tensor>]()\n  %1220 = Reshape(%1185, %1219)\n  %1221 = Constant[value = <Tensor>]()\n  %1222 = Reshape(%1192, %1221)\n  %1223 = Constant[value = <Tensor>]()\n  %1224 = Reshape(%1199, %1223)\n  %1225 = Constant[value = <Tensor>]()\n  %1226 = Reshape(%1206, %1225)\n  %1227 = Constant[value = <Tensor>]()\n  %1228 = Reshape(%1218, %1227)\n  %1229 = Add(%1220, %1222)\n  %1230 = Add(%1229, %1224)\n  %1231 = Add(%1230, %1226)\n  %1232 = Add(%1231, %1228)\n  %1233 = Shape(%1232)\n  %1234 = Shape(%1233)\n  %1235 = ConstantOfShape[value = <Tensor>](%1234)\n  %1236 = Constant[value = <Scalar Tensor []>]()\n  %1237 = Mul(%1235, %1236)\n  %1238 = Equal(%1233, %1237)\n  %1239 = Where(%1238, %1235, %1233)\n  %1240 = Expand(%1220, %1239)\n  %1241 = Unsqueeze[axes = [-1]](%1240)\n  %1242 = Shape(%1233)\n  %1243 = ConstantOfShape[value = <Tensor>](%1242)\n  %1244 = Constant[value = <Scalar Tensor []>]()\n  %1245 = Mul(%1243, %1244)\n  %1246 = Equal(%1233, %1245)\n  %1247 = Where(%1246, %1243, %1233)\n  %1248 = Expand(%1222, %1247)\n  %1249 = Unsqueeze[axes = [-1]](%1248)\n  %1250 = Shape(%1233)\n  %1251 = ConstantOfShape[value = <Tensor>](%1250)\n  %1252 = Constant[value = <Scalar Tensor []>]()\n  %1253 = Mul(%1251, %1252)\n  %1254 = Equal(%1233, %1253)\n  %1255 = Where(%1254, %1251, %1233)\n  %1256 = Expand(%1224, %1255)\n  %1257 = Unsqueeze[axes = [-1]](%1256)\n  %1258 = Shape(%1233)\n  %1259 = ConstantOfShape[value = <Tensor>](%1258)\n  %1260 = Constant[value = <Scalar Tensor []>]()\n  %1261 = Mul(%1259, %1260)\n  %1262 = Equal(%1233, %1261)\n  %1263 = Where(%1262, %1259, %1233)\n  %1264 = Expand(%1226, %1263)\n  %1265 = Unsqueeze[axes = [-1]](%1264)\n  %1266 = Shape(%1233)\n  %1267 = ConstantOfShape[value = <Tensor>](%1266)\n  %1268 = Constant[value = <Scalar Tensor []>]()\n  %1269 = Mul(%1267, %1268)\n  %1270 = Equal(%1233, %1269)\n  %1271 = Where(%1270, %1267, %1233)\n  %1272 = Expand(%1228, %1271)\n  %1273 = Unsqueeze[axes = [-1]](%1272)\n  %1274 = Concat[axis = -1](%1241, %1249, %1257, %1265, %1273)\n  %1275 = Shape(%yolo2)\n  %1276 = Constant[value = <Tensor>]()\n  %1277 = Constant[value = <Tensor>]()\n  %1278 = Constant[value = <Tensor>]()\n  %1279 = Slice(%1275, %1277, %1278, %1276)\n  %1280 = Concat[axis = 0](%1233, %1279)\n  %1281 = Reshape(%1178, %1280)\n  %1282 = ScatterND(%yolo2, %1274, %1281)\n  %1283 = Constant[value = <Tensor>]()\n  %1284 = Constant[value = <Tensor>]()\n  %1285 = Constant[value = <Tensor>]()\n  %1286 = Constant[value = <Tensor>]()\n  %1287 = Slice(%1282, %1284, %1285, %1283, %1286)\n  %1288 = Exp(%1287)\n  %1289 = Constant[value = <Tensor>]()\n  %1290 = Mul(%1288, %1289)\n  %1291 = Constant[value = <Tensor>]()\n  %1292 = Reshape(%1290, %1291)\n  %1293 = Constant[value = <Tensor>]()\n  %1295 = ConstantOfShape[value = <Tensor>](%1555)\n  %1296 = Constant[value = <Scalar Tensor []>]()\n  %1297 = Mul(%1295, %1296)\n  %1298 = Constant[value = <Tensor>]()\n  %1299 = Equal(%1298, %1297)\n  %1300 = Where(%1299, %1295, %1293)\n  %1301 = Expand(%1292, %1300)\n  %1302 = Shape(%1282)\n  %1303 = Constant[value = <Scalar Tensor []>]()\n  %1304 = Gather[axis = 0](%1302, %1303)\n  %1305 = Cast[to = 7](%1304)\n  %1306 = Constant[value = <Scalar Tensor []>]()\n  %1307 = Constant[value = <Scalar Tensor []>]()\n  %1308 = Range(%1306, %1305, %1307)\n  %1309 = Shape(%1282)\n  %1310 = Constant[value = <Scalar Tensor []>]()\n  %1311 = Gather[axis = 0](%1309, %1310)\n  %1312 = Cast[to = 7](%1311)\n  %1313 = Constant[value = <Scalar Tensor []>]()\n  %1314 = Constant[value = <Scalar Tensor []>]()\n  %1315 = Range(%1313, %1312, %1314)\n  %1316 = Shape(%1282)\n  %1317 = Constant[value = <Scalar Tensor []>]()\n  %1318 = Gather[axis = 0](%1316, %1317)\n  %1319 = Cast[to = 7](%1318)\n  %1320 = Constant[value = <Scalar Tensor []>]()\n  %1321 = Constant[value = <Scalar Tensor []>]()\n  %1322 = Range(%1320, %1319, %1321)\n  %1323 = Shape(%1282)\n  %1324 = Constant[value = <Scalar Tensor []>]()\n  %1325 = Gather[axis = 0](%1323, %1324)\n  %1326 = Cast[to = 7](%1325)\n  %1327 = Constant[value = <Scalar Tensor []>]()\n  %1328 = Constant[value = <Scalar Tensor []>]()\n  %1329 = Range(%1327, %1326, %1328)\n  %1330 = Shape(%1282)\n  %1331 = Constant[value = <Scalar Tensor []>]()\n  %1332 = Gather[axis = 0](%1330, %1331)\n  %1333 = Cast[to = 7](%1332)\n  %1334 = Constant[value = <Scalar Tensor []>]()\n  %1335 = Constant[value = <Scalar Tensor []>]()\n  %1336 = Range(%1334, %1333, %1335)\n  %1337 = Constant[value = <Tensor>]()\n  %1338 = Constant[value = <Tensor>]()\n  %1339 = Constant[value = <Tensor>]()\n  %1340 = Constant[value = <Tensor>]()\n  %1341 = Slice(%1336, %1338, %1339, %1337, %1340)\n  %1342 = Constant[value = <Tensor>]()\n  %1343 = Reshape(%1308, %1342)\n  %1344 = Constant[value = <Tensor>]()\n  %1345 = Reshape(%1315, %1344)\n  %1346 = Constant[value = <Tensor>]()\n  %1347 = Reshape(%1322, %1346)\n  %1348 = Constant[value = <Tensor>]()\n  %1349 = Reshape(%1329, %1348)\n  %1350 = Constant[value = <Tensor>]()\n  %1351 = Reshape(%1341, %1350)\n  %1352 = Add(%1343, %1345)\n  %1353 = Add(%1352, %1347)\n  %1354 = Add(%1353, %1349)\n  %1355 = Add(%1354, %1351)\n  %1356 = Shape(%1355)\n  %1357 = Shape(%1356)\n  %1358 = ConstantOfShape[value = <Tensor>](%1357)\n  %1359 = Constant[value = <Scalar Tensor []>]()\n  %1360 = Mul(%1358, %1359)\n  %1361 = Equal(%1356, %1360)\n  %1362 = Where(%1361, %1358, %1356)\n  %1363 = Expand(%1343, %1362)\n  %1364 = Unsqueeze[axes = [-1]](%1363)\n  %1365 = Shape(%1356)\n  %1366 = ConstantOfShape[value = <Tensor>](%1365)\n  %1367 = Constant[value = <Scalar Tensor []>]()\n  %1368 = Mul(%1366, %1367)\n  %1369 = Equal(%1356, %1368)\n  %1370 = Where(%1369, %1366, %1356)\n  %1371 = Expand(%1345, %1370)\n  %1372 = Unsqueeze[axes = [-1]](%1371)\n  %1373 = Shape(%1356)\n  %1374 = ConstantOfShape[value = <Tensor>](%1373)\n  %1375 = Constant[value = <Scalar Tensor []>]()\n  %1376 = Mul(%1374, %1375)\n  %1377 = Equal(%1356, %1376)\n  %1378 = Where(%1377, %1374, %1356)\n  %1379 = Expand(%1347, %1378)\n  %1380 = Unsqueeze[axes = [-1]](%1379)\n  %1381 = Shape(%1356)\n  %1382 = ConstantOfShape[value = <Tensor>](%1381)\n  %1383 = Constant[value = <Scalar Tensor []>]()\n  %1384 = Mul(%1382, %1383)\n  %1385 = Equal(%1356, %1384)\n  %1386 = Where(%1385, %1382, %1356)\n  %1387 = Expand(%1349, %1386)\n  %1388 = Unsqueeze[axes = [-1]](%1387)\n  %1389 = Shape(%1356)\n  %1390 = ConstantOfShape[value = <Tensor>](%1389)\n  %1391 = Constant[value = <Scalar Tensor []>]()\n  %1392 = Mul(%1390, %1391)\n  %1393 = Equal(%1356, %1392)\n  %1394 = Where(%1393, %1390, %1356)\n  %1395 = Expand(%1351, %1394)\n  %1396 = Unsqueeze[axes = [-1]](%1395)\n  %1397 = Concat[axis = -1](%1364, %1372, %1380, %1388, %1396)\n  %1398 = Shape(%1282)\n  %1399 = Constant[value = <Tensor>]()\n  %1400 = Constant[value = <Tensor>]()\n  %1401 = Constant[value = <Tensor>]()\n  %1402 = Slice(%1398, %1400, %1401, %1399)\n  %1403 = Concat[axis = 0](%1356, %1402)\n  %1404 = Reshape(%1301, %1403)\n  %1405 = ScatterND(%1282, %1397, %1404)\n  %1406 = Constant[value = <Tensor>]()\n  %1407 = Constant[value = <Tensor>]()\n  %1408 = Constant[value = <Tensor>]()\n  %1409 = Constant[value = <Tensor>]()\n  %1410 = Slice(%1405, %1407, %1408, %1406, %1409)\n  %1411 = Constant[value = <Scalar Tensor []>]()\n  %1412 = Mul(%1410, %1411)\n  %1413 = Constant[value = <Tensor>]()\n  %1414 = Reshape(%1412, %1413)\n  %1415 = Constant[value = <Tensor>]()\n  %1417 = ConstantOfShape[value = <Tensor>](%1556)\n  %1418 = Constant[value = <Scalar Tensor []>]()\n  %1419 = Mul(%1417, %1418)\n  %1420 = Constant[value = <Tensor>]()\n  %1421 = Equal(%1420, %1419)\n  %1422 = Where(%1421, %1417, %1415)\n  %1423 = Expand(%1414, %1422)\n  %1424 = Shape(%1405)\n  %1425 = Constant[value = <Scalar Tensor []>]()\n  %1426 = Gather[axis = 0](%1424, %1425)\n  %1427 = Cast[to = 7](%1426)\n  %1428 = Constant[value = <Scalar Tensor []>]()\n  %1429 = Constant[value = <Scalar Tensor []>]()\n  %1430 = Range(%1428, %1427, %1429)\n  %1431 = Shape(%1405)\n  %1432 = Constant[value = <Scalar Tensor []>]()\n  %1433 = Gather[axis = 0](%1431, %1432)\n  %1434 = Cast[to = 7](%1433)\n  %1435 = Constant[value = <Scalar Tensor []>]()\n  %1436 = Constant[value = <Scalar Tensor []>]()\n  %1437 = Range(%1435, %1434, %1436)\n  %1438 = Shape(%1405)\n  %1439 = Constant[value = <Scalar Tensor []>]()\n  %1440 = Gather[axis = 0](%1438, %1439)\n  %1441 = Cast[to = 7](%1440)\n  %1442 = Constant[value = <Scalar Tensor []>]()\n  %1443 = Constant[value = <Scalar Tensor []>]()\n  %1444 = Range(%1442, %1441, %1443)\n  %1445 = Shape(%1405)\n  %1446 = Constant[value = <Scalar Tensor []>]()\n  %1447 = Gather[axis = 0](%1445, %1446)\n  %1448 = Cast[to = 7](%1447)\n  %1449 = Constant[value = <Scalar Tensor []>]()\n  %1450 = Constant[value = <Scalar Tensor []>]()\n  %1451 = Range(%1449, %1448, %1450)\n  %1452 = Shape(%1405)\n  %1453 = Constant[value = <Scalar Tensor []>]()\n  %1454 = Gather[axis = 0](%1452, %1453)\n  %1455 = Cast[to = 7](%1454)\n  %1456 = Constant[value = <Scalar Tensor []>]()\n  %1457 = Constant[value = <Scalar Tensor []>]()\n  %1458 = Range(%1456, %1455, %1457)\n  %1459 = Constant[value = <Tensor>]()\n  %1460 = Constant[value = <Tensor>]()\n  %1461 = Constant[value = <Tensor>]()\n  %1462 = Constant[value = <Tensor>]()\n  %1463 = Slice(%1458, %1460, %1461, %1459, %1462)\n  %1464 = Constant[value = <Tensor>]()\n  %1465 = Reshape(%1430, %1464)\n  %1466 = Constant[value = <Tensor>]()\n  %1467 = Reshape(%1437, %1466)\n  %1468 = Constant[value = <Tensor>]()\n  %1469 = Reshape(%1444, %1468)\n  %1470 = Constant[value = <Tensor>]()\n  %1471 = Reshape(%1451, %1470)\n  %1472 = Constant[value = <Tensor>]()\n  %1473 = Reshape(%1463, %1472)\n  %1474 = Add(%1465, %1467)\n  %1475 = Add(%1474, %1469)\n  %1476 = Add(%1475, %1471)\n  %1477 = Add(%1476, %1473)\n  %1478 = Shape(%1477)\n  %1479 = Shape(%1478)\n  %1480 = ConstantOfShape[value = <Tensor>](%1479)\n  %1481 = Constant[value = <Scalar Tensor []>]()\n  %1482 = Mul(%1480, %1481)\n  %1483 = Equal(%1478, %1482)\n  %1484 = Where(%1483, %1480, %1478)\n  %1485 = Expand(%1465, %1484)\n  %1486 = Unsqueeze[axes = [-1]](%1485)\n  %1487 = Shape(%1478)\n  %1488 = ConstantOfShape[value = <Tensor>](%1487)\n  %1489 = Constant[value = <Scalar Tensor []>]()\n  %1490 = Mul(%1488, %1489)\n  %1491 = Equal(%1478, %1490)\n  %1492 = Where(%1491, %1488, %1478)\n  %1493 = Expand(%1467, %1492)\n  %1494 = Unsqueeze[axes = [-1]](%1493)\n  %1495 = Shape(%1478)\n  %1496 = ConstantOfShape[value = <Tensor>](%1495)\n  %1497 = Constant[value = <Scalar Tensor []>]()\n  %1498 = Mul(%1496, %1497)\n  %1499 = Equal(%1478, %1498)\n  %1500 = Where(%1499, %1496, %1478)\n  %1501 = Expand(%1469, %1500)\n  %1502 = Unsqueeze[axes = [-1]](%1501)\n  %1503 = Shape(%1478)\n  %1504 = ConstantOfShape[value = <Tensor>](%1503)\n  %1505 = Constant[value = <Scalar Tensor []>]()\n  %1506 = Mul(%1504, %1505)\n  %1507 = Equal(%1478, %1506)\n  %1508 = Where(%1507, %1504, %1478)\n  %1509 = Expand(%1471, %1508)\n  %1510 = Unsqueeze[axes = [-1]](%1509)\n  %1511 = Shape(%1478)\n  %1512 = ConstantOfShape[value = <Tensor>](%1511)\n  %1513 = Constant[value = <Scalar Tensor []>]()\n  %1514 = Mul(%1512, %1513)\n  %1515 = Equal(%1478, %1514)\n  %1516 = Where(%1515, %1512, %1478)\n  %1517 = Expand(%1473, %1516)\n  %1518 = Unsqueeze[axes = [-1]](%1517)\n  %1519 = Concat[axis = -1](%1486, %1494, %1502, %1510, %1518)\n  %1520 = Shape(%1405)\n  %1521 = Constant[value = <Tensor>]()\n  %1522 = Constant[value = <Tensor>]()\n  %1523 = Constant[value = <Tensor>]()\n  %1524 = Slice(%1520, %1522, %1523, %1521)\n  %1525 = Concat[axis = 0](%1478, %1524)\n  %1526 = Reshape(%1423, %1525)\n  %1527 = ScatterND(%1405, %1519, %1526)\n  %1530 = Unsqueeze[axes = [0]](%1147)\n  %1533 = Concat[axis = 0](%1530, %1557, %1558)\n  %1534 = Reshape(%1527, %1533)\n  %classes = Concat[axis = 1](%1119, %1534)\n  return %classes, %yolo1, %yolo2\n}\n"
     ]
    }
   ],
   "source": [
    "print(onnx.helper.printable_graph(onnx_model.graph))  # Print a human readable representation of the graph\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       ".BatchNorm2d.bias\"\n",
       "input: \"module_list.1.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.1.BatchNorm2d.running_var\"\n",
       "output: \"507\"\n",
       "name: \"BatchNormalization_4\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"507\"\n",
       "output: \"508\"\n",
       "name: \"LeakyRelu_5\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"508\"\n",
       "input: \"module_list.2.Conv2d.weight\"\n",
       "output: \"509\"\n",
       "name: \"Conv_6\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 8\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"509\"\n",
       "input: \"module_list.2.BatchNorm2d.weight\"\n",
       "input: \"module_list.2.BatchNorm2d.bias\"\n",
       "input: \"module_list.2.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.2.BatchNorm2d.running_var\"\n",
       "output: \"510\"\n",
       "name: \"BatchNormalization_7\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"510\"\n",
       "output: \"511\"\n",
       "name: \"LeakyRelu_8\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"511\"\n",
       "input: \"module_list.3.Conv2d.weight\"\n",
       "output: \"512\"\n",
       "name: \"Conv_9\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"512\"\n",
       "input: \"module_list.3.BatchNorm2d.weight\"\n",
       "input: \"module_list.3.BatchNorm2d.bias\"\n",
       "input: \"module_list.3.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.3.BatchNorm2d.running_var\"\n",
       "output: \"513\"\n",
       "name: \"BatchNormalization_10\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"513\"\n",
       "input: \"module_list.4.Conv2d.weight\"\n",
       "output: \"514\"\n",
       "name: \"Conv_11\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"514\"\n",
       "input: \"module_list.4.BatchNorm2d.weight\"\n",
       "input: \"module_list.4.BatchNorm2d.bias\"\n",
       "input: \"module_list.4.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.4.BatchNorm2d.running_var\"\n",
       "output: \"515\"\n",
       "name: \"BatchNormalization_12\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"515\"\n",
       "output: \"516\"\n",
       "name: \"LeakyRelu_13\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"516\"\n",
       "input: \"module_list.5.Conv2d.weight\"\n",
       "output: \"517\"\n",
       "name: \"Conv_14\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 8\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"517\"\n",
       "input: \"module_list.5.BatchNorm2d.weight\"\n",
       "input: \"module_list.5.BatchNorm2d.bias\"\n",
       "input: \"module_list.5.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.5.BatchNorm2d.running_var\"\n",
       "output: \"518\"\n",
       "name: \"BatchNormalization_15\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"518\"\n",
       "output: \"519\"\n",
       "name: \"LeakyRelu_16\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"519\"\n",
       "input: \"module_list.6.Conv2d.weight\"\n",
       "output: \"520\"\n",
       "name: \"Conv_17\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"520\"\n",
       "input: \"module_list.6.BatchNorm2d.weight\"\n",
       "input: \"module_list.6.BatchNorm2d.bias\"\n",
       "input: \"module_list.6.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.6.BatchNorm2d.running_var\"\n",
       "output: \"521\"\n",
       "name: \"BatchNormalization_18\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"521\"\n",
       "input: \"513\"\n",
       "output: \"522\"\n",
       "name: \"Add_19\"\n",
       "op_type: \"Add\"\n",
       ", input: \"522\"\n",
       "input: \"module_list.9.Conv2d.weight\"\n",
       "output: \"523\"\n",
       "name: \"Conv_20\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"523\"\n",
       "input: \"module_list.9.BatchNorm2d.weight\"\n",
       "input: \"module_list.9.BatchNorm2d.bias\"\n",
       "input: \"module_list.9.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.9.BatchNorm2d.running_var\"\n",
       "output: \"524\"\n",
       "name: \"BatchNormalization_21\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"524\"\n",
       "output: \"525\"\n",
       "name: \"LeakyRelu_22\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"525\"\n",
       "input: \"module_list.10.Conv2d.weight\"\n",
       "output: \"526\"\n",
       "name: \"Conv_23\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 24\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"526\"\n",
       "input: \"module_list.10.BatchNorm2d.weight\"\n",
       "input: \"module_list.10.BatchNorm2d.bias\"\n",
       "input: \"module_list.10.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.10.BatchNorm2d.running_var\"\n",
       "output: \"527\"\n",
       "name: \"BatchNormalization_24\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"527\"\n",
       "output: \"528\"\n",
       "name: \"LeakyRelu_25\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"528\"\n",
       "input: \"module_list.11.Conv2d.weight\"\n",
       "output: \"529\"\n",
       "name: \"Conv_26\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"529\"\n",
       "input: \"module_list.11.BatchNorm2d.weight\"\n",
       "input: \"module_list.11.BatchNorm2d.bias\"\n",
       "input: \"module_list.11.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.11.BatchNorm2d.running_var\"\n",
       "output: \"530\"\n",
       "name: \"BatchNormalization_27\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"530\"\n",
       "input: \"module_list.12.Conv2d.weight\"\n",
       "output: \"531\"\n",
       "name: \"Conv_28\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"531\"\n",
       "input: \"module_list.12.BatchNorm2d.weight\"\n",
       "input: \"module_list.12.BatchNorm2d.bias\"\n",
       "input: \"module_list.12.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.12.BatchNorm2d.running_var\"\n",
       "output: \"532\"\n",
       "name: \"BatchNormalization_29\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"532\"\n",
       "output: \"533\"\n",
       "name: \"LeakyRelu_30\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"533\"\n",
       "input: \"module_list.13.Conv2d.weight\"\n",
       "output: \"534\"\n",
       "name: \"Conv_31\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 32\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"534\"\n",
       "input: \"module_list.13.BatchNorm2d.weight\"\n",
       "input: \"module_list.13.BatchNorm2d.bias\"\n",
       "input: \"module_list.13.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.13.BatchNorm2d.running_var\"\n",
       "output: \"535\"\n",
       "name: \"BatchNormalization_32\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"535\"\n",
       "output: \"536\"\n",
       "name: \"LeakyRelu_33\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"536\"\n",
       "input: \"module_list.14.Conv2d.weight\"\n",
       "output: \"537\"\n",
       "name: \"Conv_34\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"537\"\n",
       "input: \"module_list.14.BatchNorm2d.weight\"\n",
       "input: \"module_list.14.BatchNorm2d.bias\"\n",
       "input: \"module_list.14.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.14.BatchNorm2d.running_var\"\n",
       "output: \"538\"\n",
       "name: \"BatchNormalization_35\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"538\"\n",
       "input: \"530\"\n",
       "output: \"539\"\n",
       "name: \"Add_36\"\n",
       "op_type: \"Add\"\n",
       ", input: \"539\"\n",
       "input: \"module_list.17.Conv2d.weight\"\n",
       "output: \"540\"\n",
       "name: \"Conv_37\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"540\"\n",
       "input: \"module_list.17.BatchNorm2d.weight\"\n",
       "input: \"module_list.17.BatchNorm2d.bias\"\n",
       "input: \"module_list.17.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.17.BatchNorm2d.running_var\"\n",
       "output: \"541\"\n",
       "name: \"BatchNormalization_38\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"541\"\n",
       "output: \"542\"\n",
       "name: \"LeakyRelu_39\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"542\"\n",
       "input: \"module_list.18.Conv2d.weight\"\n",
       "output: \"543\"\n",
       "name: \"Conv_40\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 32\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"543\"\n",
       "input: \"module_list.18.BatchNorm2d.weight\"\n",
       "input: \"module_list.18.BatchNorm2d.bias\"\n",
       "input: \"module_list.18.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.18.BatchNorm2d.running_var\"\n",
       "output: \"544\"\n",
       "name: \"BatchNormalization_41\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"544\"\n",
       "output: \"545\"\n",
       "name: \"LeakyRelu_42\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"545\"\n",
       "input: \"module_list.19.Conv2d.weight\"\n",
       "output: \"546\"\n",
       "name: \"Conv_43\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"546\"\n",
       "input: \"module_list.19.BatchNorm2d.weight\"\n",
       "input: \"module_list.19.BatchNorm2d.bias\"\n",
       "input: \"module_list.19.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.19.BatchNorm2d.running_var\"\n",
       "output: \"547\"\n",
       "name: \"BatchNormalization_44\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"547\"\n",
       "input: \"539\"\n",
       "output: \"548\"\n",
       "name: \"Add_45\"\n",
       "op_type: \"Add\"\n",
       ", input: \"548\"\n",
       "input: \"module_list.22.Conv2d.weight\"\n",
       "output: \"549\"\n",
       "name: \"Conv_46\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"549\"\n",
       "input: \"module_list.22.BatchNorm2d.weight\"\n",
       "input: \"module_list.22.BatchNorm2d.bias\"\n",
       "input: \"module_list.22.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.22.BatchNorm2d.running_var\"\n",
       "output: \"550\"\n",
       "name: \"BatchNormalization_47\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"550\"\n",
       "output: \"551\"\n",
       "name: \"LeakyRelu_48\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"551\"\n",
       "input: \"module_list.23.Conv2d.weight\"\n",
       "output: \"552\"\n",
       "name: \"Conv_49\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 32\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"552\"\n",
       "input: \"module_list.23.BatchNorm2d.weight\"\n",
       "input: \"module_list.23.BatchNorm2d.bias\"\n",
       "input: \"module_list.23.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.23.BatchNorm2d.running_var\"\n",
       "output: \"553\"\n",
       "name: \"BatchNormalization_50\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"553\"\n",
       "output: \"554\"\n",
       "name: \"LeakyRelu_51\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"554\"\n",
       "input: \"module_list.24.Conv2d.weight\"\n",
       "output: \"555\"\n",
       "name: \"Conv_52\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"555\"\n",
       "input: \"module_list.24.BatchNorm2d.weight\"\n",
       "input: \"module_list.24.BatchNorm2d.bias\"\n",
       "input: \"module_list.24.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.24.BatchNorm2d.running_var\"\n",
       "output: \"556\"\n",
       "name: \"BatchNormalization_53\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"556\"\n",
       "input: \"module_list.25.Conv2d.weight\"\n",
       "output: \"557\"\n",
       "name: \"Conv_54\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"557\"\n",
       "input: \"module_list.25.BatchNorm2d.weight\"\n",
       "input: \"module_list.25.BatchNorm2d.bias\"\n",
       "input: \"module_list.25.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.25.BatchNorm2d.running_var\"\n",
       "output: \"558\"\n",
       "name: \"BatchNormalization_55\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"558\"\n",
       "output: \"559\"\n",
       "name: \"LeakyRelu_56\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"559\"\n",
       "input: \"module_list.26.Conv2d.weight\"\n",
       "output: \"560\"\n",
       "name: \"Conv_57\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 48\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"560\"\n",
       "input: \"module_list.26.BatchNorm2d.weight\"\n",
       "input: \"module_list.26.BatchNorm2d.bias\"\n",
       "input: \"module_list.26.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.26.BatchNorm2d.running_var\"\n",
       "output: \"561\"\n",
       "name: \"BatchNormalization_58\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"561\"\n",
       "output: \"562\"\n",
       "name: \"LeakyRelu_59\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"562\"\n",
       "input: \"module_list.27.Conv2d.weight\"\n",
       "output: \"563\"\n",
       "name: \"Conv_60\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"563\"\n",
       "input: \"module_list.27.BatchNorm2d.weight\"\n",
       "input: \"module_list.27.BatchNorm2d.bias\"\n",
       "input: \"module_list.27.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.27.BatchNorm2d.running_var\"\n",
       "output: \"564\"\n",
       "name: \"BatchNormalization_61\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"564\"\n",
       "input: \"556\"\n",
       "output: \"565\"\n",
       "name: \"Add_62\"\n",
       "op_type: \"Add\"\n",
       ", input: \"565\"\n",
       "input: \"module_list.30.Conv2d.weight\"\n",
       "output: \"566\"\n",
       "name: \"Conv_63\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"566\"\n",
       "input: \"module_list.30.BatchNorm2d.weight\"\n",
       "input: \"module_list.30.BatchNorm2d.bias\"\n",
       "input: \"module_list.30.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.30.BatchNorm2d.running_var\"\n",
       "output: \"567\"\n",
       "name: \"BatchNormalization_64\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"567\"\n",
       "output: \"568\"\n",
       "name: \"LeakyRelu_65\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"568\"\n",
       "input: \"module_list.31.Conv2d.weight\"\n",
       "output: \"569\"\n",
       "name: \"Conv_66\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 48\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"569\"\n",
       "input: \"module_list.31.BatchNorm2d.weight\"\n",
       "input: \"module_list.31.BatchNorm2d.bias\"\n",
       "input: \"module_list.31.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.31.BatchNorm2d.running_var\"\n",
       "output: \"570\"\n",
       "name: \"BatchNormalization_67\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"570\"\n",
       "output: \"571\"\n",
       "name: \"LeakyRelu_68\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"571\"\n",
       "input: \"module_list.32.Conv2d.weight\"\n",
       "output: \"572\"\n",
       "name: \"Conv_69\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"572\"\n",
       "input: \"module_list.32.BatchNorm2d.weight\"\n",
       "input: \"module_list.32.BatchNorm2d.bias\"\n",
       "input: \"module_list.32.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.32.BatchNorm2d.running_var\"\n",
       "output: \"573\"\n",
       "name: \"BatchNormalization_70\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"573\"\n",
       "input: \"565\"\n",
       "output: \"574\"\n",
       "name: \"Add_71\"\n",
       "op_type: \"Add\"\n",
       ", input: \"574\"\n",
       "input: \"module_list.35.Conv2d.weight\"\n",
       "output: \"575\"\n",
       "name: \"Conv_72\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"575\"\n",
       "input: \"module_list.35.BatchNorm2d.weight\"\n",
       "input: \"module_list.35.BatchNorm2d.bias\"\n",
       "input: \"module_list.35.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.35.BatchNorm2d.running_var\"\n",
       "output: \"576\"\n",
       "name: \"BatchNormalization_73\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"576\"\n",
       "output: \"577\"\n",
       "name: \"LeakyRelu_74\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"577\"\n",
       "input: \"module_list.36.Conv2d.weight\"\n",
       "output: \"578\"\n",
       "name: \"Conv_75\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 48\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"578\"\n",
       "input: \"module_list.36.BatchNorm2d.weight\"\n",
       "input: \"module_list.36.BatchNorm2d.bias\"\n",
       "input: \"module_list.36.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.36.BatchNorm2d.running_var\"\n",
       "output: \"579\"\n",
       "name: \"BatchNormalization_76\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"579\"\n",
       "output: \"580\"\n",
       "name: \"LeakyRelu_77\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"580\"\n",
       "input: \"module_list.37.Conv2d.weight\"\n",
       "output: \"581\"\n",
       "name: \"Conv_78\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"581\"\n",
       "input: \"module_list.37.BatchNorm2d.weight\"\n",
       "input: \"module_list.37.BatchNorm2d.bias\"\n",
       "input: \"module_list.37.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.37.BatchNorm2d.running_var\"\n",
       "output: \"582\"\n",
       "name: \"BatchNormalization_79\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"582\"\n",
       "input: \"module_list.38.Conv2d.weight\"\n",
       "output: \"583\"\n",
       "name: \"Conv_80\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"583\"\n",
       "input: \"module_list.38.BatchNorm2d.weight\"\n",
       "input: \"module_list.38.BatchNorm2d.bias\"\n",
       "input: \"module_list.38.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.38.BatchNorm2d.running_var\"\n",
       "output: \"584\"\n",
       "name: \"BatchNormalization_81\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"584\"\n",
       "output: \"585\"\n",
       "name: \"LeakyRelu_82\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"585\"\n",
       "input: \"module_list.39.Conv2d.weight\"\n",
       "output: \"586\"\n",
       "name: \"Conv_83\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"586\"\n",
       "input: \"module_list.39.BatchNorm2d.weight\"\n",
       "input: \"module_list.39.BatchNorm2d.bias\"\n",
       "input: \"module_list.39.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.39.BatchNorm2d.running_var\"\n",
       "output: \"587\"\n",
       "name: \"BatchNormalization_84\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"587\"\n",
       "output: \"588\"\n",
       "name: \"LeakyRelu_85\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"588\"\n",
       "input: \"module_list.40.Conv2d.weight\"\n",
       "output: \"589\"\n",
       "name: \"Conv_86\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"589\"\n",
       "input: \"module_list.40.BatchNorm2d.weight\"\n",
       "input: \"module_list.40.BatchNorm2d.bias\"\n",
       "input: \"module_list.40.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.40.BatchNorm2d.running_var\"\n",
       "output: \"590\"\n",
       "name: \"BatchNormalization_87\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"590\"\n",
       "input: \"582\"\n",
       "output: \"591\"\n",
       "name: \"Add_88\"\n",
       "op_type: \"Add\"\n",
       ", input: \"591\"\n",
       "input: \"module_list.43.Conv2d.weight\"\n",
       "output: \"592\"\n",
       "name: \"Conv_89\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"592\"\n",
       "input: \"module_list.43.BatchNorm2d.weight\"\n",
       "input: \"module_list.43.BatchNorm2d.bias\"\n",
       "input: \"module_list.43.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.43.BatchNorm2d.running_var\"\n",
       "output: \"593\"\n",
       "name: \"BatchNormalization_90\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"593\"\n",
       "output: \"594\"\n",
       "name: \"LeakyRelu_91\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"594\"\n",
       "input: \"module_list.44.Conv2d.weight\"\n",
       "output: \"595\"\n",
       "name: \"Conv_92\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"595\"\n",
       "input: \"module_list.44.BatchNorm2d.weight\"\n",
       "input: \"module_list.44.BatchNorm2d.bias\"\n",
       "input: \"module_list.44.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.44.BatchNorm2d.running_var\"\n",
       "output: \"596\"\n",
       "name: \"BatchNormalization_93\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"596\"\n",
       "output: \"597\"\n",
       "name: \"LeakyRelu_94\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"597\"\n",
       "input: \"module_list.45.Conv2d.weight\"\n",
       "output: \"598\"\n",
       "name: \"Conv_95\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"598\"\n",
       "input: \"module_list.45.BatchNorm2d.weight\"\n",
       "input: \"module_list.45.BatchNorm2d.bias\"\n",
       "input: \"module_list.45.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.45.BatchNorm2d.running_var\"\n",
       "output: \"599\"\n",
       "name: \"BatchNormalization_96\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"599\"\n",
       "input: \"591\"\n",
       "output: \"600\"\n",
       "name: \"Add_97\"\n",
       "op_type: \"Add\"\n",
       ", input: \"600\"\n",
       "input: \"module_list.48.Conv2d.weight\"\n",
       "output: \"601\"\n",
       "name: \"Conv_98\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"601\"\n",
       "input: \"module_list.48.BatchNorm2d.weight\"\n",
       "input: \"module_list.48.BatchNorm2d.bias\"\n",
       "input: \"module_list.48.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.48.BatchNorm2d.running_var\"\n",
       "output: \"602\"\n",
       "name: \"BatchNormalization_99\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"602\"\n",
       "output: \"603\"\n",
       "name: \"LeakyRelu_100\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"603\"\n",
       "input: \"module_list.49.Conv2d.weight\"\n",
       "output: \"604\"\n",
       "name: \"Conv_101\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"604\"\n",
       "input: \"module_list.49.BatchNorm2d.weight\"\n",
       "input: \"module_list.49.BatchNorm2d.bias\"\n",
       "input: \"module_list.49.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.49.BatchNorm2d.running_var\"\n",
       "output: \"605\"\n",
       "name: \"BatchNormalization_102\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"605\"\n",
       "output: \"606\"\n",
       "name: \"LeakyRelu_103\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"606\"\n",
       "input: \"module_list.50.Conv2d.weight\"\n",
       "output: \"607\"\n",
       "name: \"Conv_104\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"607\"\n",
       "input: \"module_list.50.BatchNorm2d.weight\"\n",
       "input: \"module_list.50.BatchNorm2d.bias\"\n",
       "input: \"module_list.50.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.50.BatchNorm2d.running_var\"\n",
       "output: \"608\"\n",
       "name: \"BatchNormalization_105\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"608\"\n",
       "input: \"600\"\n",
       "output: \"609\"\n",
       "name: \"Add_106\"\n",
       "op_type: \"Add\"\n",
       ", input: \"609\"\n",
       "input: \"module_list.53.Conv2d.weight\"\n",
       "output: \"610\"\n",
       "name: \"Conv_107\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"610\"\n",
       "input: \"module_list.53.BatchNorm2d.weight\"\n",
       "input: \"module_list.53.BatchNorm2d.bias\"\n",
       "input: \"module_list.53.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.53.BatchNorm2d.running_var\"\n",
       "output: \"611\"\n",
       "name: \"BatchNormalization_108\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"611\"\n",
       "output: \"612\"\n",
       "name: \"LeakyRelu_109\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"612\"\n",
       "input: \"module_list.54.Conv2d.weight\"\n",
       "output: \"613\"\n",
       "name: \"Conv_110\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"613\"\n",
       "input: \"module_list.54.BatchNorm2d.weight\"\n",
       "input: \"module_list.54.BatchNorm2d.bias\"\n",
       "input: \"module_list.54.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.54.BatchNorm2d.running_var\"\n",
       "output: \"614\"\n",
       "name: \"BatchNormalization_111\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"614\"\n",
       "output: \"615\"\n",
       "name: \"LeakyRelu_112\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"615\"\n",
       "input: \"module_list.55.Conv2d.weight\"\n",
       "output: \"616\"\n",
       "name: \"Conv_113\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"616\"\n",
       "input: \"module_list.55.BatchNorm2d.weight\"\n",
       "input: \"module_list.55.BatchNorm2d.bias\"\n",
       "input: \"module_list.55.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.55.BatchNorm2d.running_var\"\n",
       "output: \"617\"\n",
       "name: \"BatchNormalization_114\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"617\"\n",
       "input: \"609\"\n",
       "output: \"618\"\n",
       "name: \"Add_115\"\n",
       "op_type: \"Add\"\n",
       ", input: \"618\"\n",
       "input: \"module_list.58.Conv2d.weight\"\n",
       "output: \"619\"\n",
       "name: \"Conv_116\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"619\"\n",
       "input: \"module_list.58.BatchNorm2d.weight\"\n",
       "input: \"module_list.58.BatchNorm2d.bias\"\n",
       "input: \"module_list.58.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.58.BatchNorm2d.running_var\"\n",
       "output: \"620\"\n",
       "name: \"BatchNormalization_117\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"620\"\n",
       "output: \"621\"\n",
       "name: \"LeakyRelu_118\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"621\"\n",
       "input: \"module_list.59.Conv2d.weight\"\n",
       "output: \"622\"\n",
       "name: \"Conv_119\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"622\"\n",
       "input: \"module_list.59.BatchNorm2d.weight\"\n",
       "input: \"module_list.59.BatchNorm2d.bias\"\n",
       "input: \"module_list.59.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.59.BatchNorm2d.running_var\"\n",
       "output: \"623\"\n",
       "name: \"BatchNormalization_120\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"623\"\n",
       "output: \"624\"\n",
       "name: \"LeakyRelu_121\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"624\"\n",
       "input: \"module_list.60.Conv2d.weight\"\n",
       "output: \"625\"\n",
       "name: \"Conv_122\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"625\"\n",
       "input: \"module_list.60.BatchNorm2d.weight\"\n",
       "input: \"module_list.60.BatchNorm2d.bias\"\n",
       "input: \"module_list.60.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.60.BatchNorm2d.running_var\"\n",
       "output: \"626\"\n",
       "name: \"BatchNormalization_123\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"626\"\n",
       "input: \"module_list.61.Conv2d.weight\"\n",
       "output: \"627\"\n",
       "name: \"Conv_124\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"627\"\n",
       "input: \"module_list.61.BatchNorm2d.weight\"\n",
       "input: \"module_list.61.BatchNorm2d.bias\"\n",
       "input: \"module_list.61.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.61.BatchNorm2d.running_var\"\n",
       "output: \"628\"\n",
       "name: \"BatchNormalization_125\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"628\"\n",
       "output: \"629\"\n",
       "name: \"LeakyRelu_126\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"629\"\n",
       "input: \"module_list.62.Conv2d.weight\"\n",
       "output: \"630\"\n",
       "name: \"Conv_127\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 136\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"630\"\n",
       "input: \"module_list.62.BatchNorm2d.weight\"\n",
       "input: \"module_list.62.BatchNorm2d.bias\"\n",
       "input: \"module_list.62.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.62.BatchNorm2d.running_var\"\n",
       "output: \"631\"\n",
       "name: \"BatchNormalization_128\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"631\"\n",
       "output: \"632\"\n",
       "name: \"LeakyRelu_129\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"632\"\n",
       "input: \"module_list.63.Conv2d.weight\"\n",
       "output: \"633\"\n",
       "name: \"Conv_130\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"633\"\n",
       "input: \"module_list.63.BatchNorm2d.weight\"\n",
       "input: \"module_list.63.BatchNorm2d.bias\"\n",
       "input: \"module_list.63.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.63.BatchNorm2d.running_var\"\n",
       "output: \"634\"\n",
       "name: \"BatchNormalization_131\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"634\"\n",
       "input: \"626\"\n",
       "output: \"635\"\n",
       "name: \"Add_132\"\n",
       "op_type: \"Add\"\n",
       ", input: \"635\"\n",
       "input: \"module_list.66.Conv2d.weight\"\n",
       "output: \"636\"\n",
       "name: \"Conv_133\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"636\"\n",
       "input: \"module_list.66.BatchNorm2d.weight\"\n",
       "input: \"module_list.66.BatchNorm2d.bias\"\n",
       "input: \"module_list.66.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.66.BatchNorm2d.running_var\"\n",
       "output: \"637\"\n",
       "name: \"BatchNormalization_134\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"637\"\n",
       "output: \"638\"\n",
       "name: \"LeakyRelu_135\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"638\"\n",
       "input: \"module_list.67.Conv2d.weight\"\n",
       "output: \"639\"\n",
       "name: \"Conv_136\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 136\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"639\"\n",
       "input: \"module_list.67.BatchNorm2d.weight\"\n",
       "input: \"module_list.67.BatchNorm2d.bias\"\n",
       "input: \"module_list.67.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.67.BatchNorm2d.running_var\"\n",
       "output: \"640\"\n",
       "name: \"BatchNormalization_137\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"640\"\n",
       "output: \"641\"\n",
       "name: \"LeakyRelu_138\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"641\"\n",
       "input: \"module_list.68.Conv2d.weight\"\n",
       "output: \"642\"\n",
       "name: \"Conv_139\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"642\"\n",
       "input: \"module_list.68.BatchNorm2d.weight\"\n",
       "input: \"module_list.68.BatchNorm2d.bias\"\n",
       "input: \"module_list.68.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.68.BatchNorm2d.running_var\"\n",
       "output: \"643\"\n",
       "name: \"BatchNormalization_140\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"643\"\n",
       "input: \"635\"\n",
       "output: \"644\"\n",
       "name: \"Add_141\"\n",
       "op_type: \"Add\"\n",
       ", input: \"644\"\n",
       "input: \"module_list.71.Conv2d.weight\"\n",
       "output: \"645\"\n",
       "name: \"Conv_142\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"645\"\n",
       "input: \"module_list.71.BatchNorm2d.weight\"\n",
       "input: \"module_list.71.BatchNorm2d.bias\"\n",
       "input: \"module_list.71.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.71.BatchNorm2d.running_var\"\n",
       "output: \"646\"\n",
       "name: \"BatchNormalization_143\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"646\"\n",
       "output: \"647\"\n",
       "name: \"LeakyRelu_144\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"647\"\n",
       "input: \"module_list.72.Conv2d.weight\"\n",
       "output: \"648\"\n",
       "name: \"Conv_145\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 136\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"648\"\n",
       "input: \"module_list.72.BatchNorm2d.weight\"\n",
       "input: \"module_list.72.BatchNorm2d.bias\"\n",
       "input: \"module_list.72.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.72.BatchNorm2d.running_var\"\n",
       "output: \"649\"\n",
       "name: \"BatchNormalization_146\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"649\"\n",
       "output: \"650\"\n",
       "name: \"LeakyRelu_147\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"650\"\n",
       "input: \"module_list.73.Conv2d.weight\"\n",
       "output: \"651\"\n",
       "name: \"Conv_148\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"651\"\n",
       "input: \"module_list.73.BatchNorm2d.weight\"\n",
       "input: \"module_list.73.BatchNorm2d.bias\"\n",
       "input: \"module_list.73.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.73.BatchNorm2d.running_var\"\n",
       "output: \"652\"\n",
       "name: \"BatchNormalization_149\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"652\"\n",
       "input: \"644\"\n",
       "output: \"653\"\n",
       "name: \"Add_150\"\n",
       "op_type: \"Add\"\n",
       ", input: \"653\"\n",
       "input: \"module_list.76.Conv2d.weight\"\n",
       "output: \"654\"\n",
       "name: \"Conv_151\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"654\"\n",
       "input: \"module_list.76.BatchNorm2d.weight\"\n",
       "input: \"module_list.76.BatchNorm2d.bias\"\n",
       "input: \"module_list.76.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.76.BatchNorm2d.running_var\"\n",
       "output: \"655\"\n",
       "name: \"BatchNormalization_152\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"655\"\n",
       "output: \"656\"\n",
       "name: \"LeakyRelu_153\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"656\"\n",
       "input: \"module_list.77.Conv2d.weight\"\n",
       "output: \"657\"\n",
       "name: \"Conv_154\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 136\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"657\"\n",
       "input: \"module_list.77.BatchNorm2d.weight\"\n",
       "input: \"module_list.77.BatchNorm2d.bias\"\n",
       "input: \"module_list.77.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.77.BatchNorm2d.running_var\"\n",
       "output: \"658\"\n",
       "name: \"BatchNormalization_155\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"658\"\n",
       "output: \"659\"\n",
       "name: \"LeakyRelu_156\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"659\"\n",
       "input: \"module_list.78.Conv2d.weight\"\n",
       "output: \"660\"\n",
       "name: \"Conv_157\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"660\"\n",
       "input: \"module_list.78.BatchNorm2d.weight\"\n",
       "input: \"module_list.78.BatchNorm2d.bias\"\n",
       "input: \"module_list.78.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.78.BatchNorm2d.running_var\"\n",
       "output: \"661\"\n",
       "name: \"BatchNormalization_158\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"661\"\n",
       "input: \"653\"\n",
       "output: \"662\"\n",
       "name: \"Add_159\"\n",
       "op_type: \"Add\"\n",
       ", input: \"662\"\n",
       "input: \"module_list.81.Conv2d.weight\"\n",
       "output: \"663\"\n",
       "name: \"Conv_160\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"663\"\n",
       "input: \"module_list.81.BatchNorm2d.weight\"\n",
       "input: \"module_list.81.BatchNorm2d.bias\"\n",
       "input: \"module_list.81.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.81.BatchNorm2d.running_var\"\n",
       "output: \"664\"\n",
       "name: \"BatchNormalization_161\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"664\"\n",
       "output: \"665\"\n",
       "name: \"LeakyRelu_162\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"665\"\n",
       "input: \"module_list.82.Conv2d.weight\"\n",
       "output: \"666\"\n",
       "name: \"Conv_163\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 136\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"666\"\n",
       "input: \"module_list.82.BatchNorm2d.weight\"\n",
       "input: \"module_list.82.BatchNorm2d.bias\"\n",
       "input: \"module_list.82.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.82.BatchNorm2d.running_var\"\n",
       "output: \"667\"\n",
       "name: \"BatchNormalization_164\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"667\"\n",
       "output: \"668\"\n",
       "name: \"LeakyRelu_165\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"668\"\n",
       "input: \"module_list.83.Conv2d.weight\"\n",
       "output: \"669\"\n",
       "name: \"Conv_166\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"669\"\n",
       "input: \"module_list.83.BatchNorm2d.weight\"\n",
       "input: \"module_list.83.BatchNorm2d.bias\"\n",
       "input: \"module_list.83.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.83.BatchNorm2d.running_var\"\n",
       "output: \"670\"\n",
       "name: \"BatchNormalization_167\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"670\"\n",
       "input: \"module_list.84.Conv2d.weight\"\n",
       "output: \"671\"\n",
       "name: \"Conv_168\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"671\"\n",
       "input: \"module_list.84.BatchNorm2d.weight\"\n",
       "input: \"module_list.84.BatchNorm2d.bias\"\n",
       "input: \"module_list.84.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.84.BatchNorm2d.running_var\"\n",
       "output: \"672\"\n",
       "name: \"BatchNormalization_169\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"672\"\n",
       "output: \"673\"\n",
       "name: \"LeakyRelu_170\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"673\"\n",
       "input: \"module_list.85.Conv2d.weight\"\n",
       "output: \"674\"\n",
       "name: \"Conv_171\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 224\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"674\"\n",
       "input: \"module_list.85.BatchNorm2d.weight\"\n",
       "input: \"module_list.85.BatchNorm2d.bias\"\n",
       "input: \"module_list.85.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.85.BatchNorm2d.running_var\"\n",
       "output: \"675\"\n",
       "name: \"BatchNormalization_172\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"675\"\n",
       "output: \"676\"\n",
       "name: \"LeakyRelu_173\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"676\"\n",
       "input: \"module_list.86.Conv2d.weight\"\n",
       "output: \"677\"\n",
       "name: \"Conv_174\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"677\"\n",
       "input: \"module_list.86.BatchNorm2d.weight\"\n",
       "input: \"module_list.86.BatchNorm2d.bias\"\n",
       "input: \"module_list.86.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.86.BatchNorm2d.running_var\"\n",
       "output: \"678\"\n",
       "name: \"BatchNormalization_175\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"678\"\n",
       "input: \"670\"\n",
       "output: \"679\"\n",
       "name: \"Add_176\"\n",
       "op_type: \"Add\"\n",
       ", input: \"679\"\n",
       "input: \"module_list.89.Conv2d.weight\"\n",
       "output: \"680\"\n",
       "name: \"Conv_177\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"680\"\n",
       "input: \"module_list.89.BatchNorm2d.weight\"\n",
       "input: \"module_list.89.BatchNorm2d.bias\"\n",
       "input: \"module_list.89.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.89.BatchNorm2d.running_var\"\n",
       "output: \"681\"\n",
       "name: \"BatchNormalization_178\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"681\"\n",
       "output: \"682\"\n",
       "name: \"LeakyRelu_179\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"682\"\n",
       "input: \"module_list.90.Conv2d.weight\"\n",
       "output: \"683\"\n",
       "name: \"Conv_180\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 224\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"683\"\n",
       "input: \"module_list.90.BatchNorm2d.weight\"\n",
       "input: \"module_list.90.BatchNorm2d.bias\"\n",
       "input: \"module_list.90.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.90.BatchNorm2d.running_var\"\n",
       "output: \"684\"\n",
       "name: \"BatchNormalization_181\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"684\"\n",
       "output: \"685\"\n",
       "name: \"LeakyRelu_182\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"685\"\n",
       "input: \"module_list.91.Conv2d.weight\"\n",
       "output: \"686\"\n",
       "name: \"Conv_183\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"686\"\n",
       "input: \"module_list.91.BatchNorm2d.weight\"\n",
       "input: \"module_list.91.BatchNorm2d.bias\"\n",
       "input: \"module_list.91.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.91.BatchNorm2d.running_var\"\n",
       "output: \"687\"\n",
       "name: \"BatchNormalization_184\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"687\"\n",
       "input: \"679\"\n",
       "output: \"688\"\n",
       "name: \"Add_185\"\n",
       "op_type: \"Add\"\n",
       ", input: \"688\"\n",
       "input: \"module_list.94.Conv2d.weight\"\n",
       "output: \"689\"\n",
       "name: \"Conv_186\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"689\"\n",
       "input: \"module_list.94.BatchNorm2d.weight\"\n",
       "input: \"module_list.94.BatchNorm2d.bias\"\n",
       "input: \"module_list.94.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.94.BatchNorm2d.running_var\"\n",
       "output: \"690\"\n",
       "name: \"BatchNormalization_187\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"690\"\n",
       "output: \"691\"\n",
       "name: \"LeakyRelu_188\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"691\"\n",
       "input: \"module_list.95.Conv2d.weight\"\n",
       "output: \"692\"\n",
       "name: \"Conv_189\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 224\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"692\"\n",
       "input: \"module_list.95.BatchNorm2d.weight\"\n",
       "input: \"module_list.95.BatchNorm2d.bias\"\n",
       "input: \"module_list.95.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.95.BatchNorm2d.running_var\"\n",
       "output: \"693\"\n",
       "name: \"BatchNormalization_190\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"693\"\n",
       "output: \"694\"\n",
       "name: \"LeakyRelu_191\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"694\"\n",
       "input: \"module_list.96.Conv2d.weight\"\n",
       "output: \"695\"\n",
       "name: \"Conv_192\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"695\"\n",
       "input: \"module_list.96.BatchNorm2d.weight\"\n",
       "input: \"module_list.96.BatchNorm2d.bias\"\n",
       "input: \"module_list.96.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.96.BatchNorm2d.running_var\"\n",
       "output: \"696\"\n",
       "name: \"BatchNormalization_193\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"696\"\n",
       "input: \"688\"\n",
       "output: \"697\"\n",
       "name: \"Add_194\"\n",
       "op_type: \"Add\"\n",
       ", input: \"697\"\n",
       "input: \"module_list.99.Conv2d.weight\"\n",
       "output: \"698\"\n",
       "name: \"Conv_195\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"698\"\n",
       "input: \"module_list.99.BatchNorm2d.weight\"\n",
       "input: \"module_list.99.BatchNorm2d.bias\"\n",
       "input: \"module_list.99.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.99.BatchNorm2d.running_var\"\n",
       "output: \"699\"\n",
       "name: \"BatchNormalization_196\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"699\"\n",
       "output: \"700\"\n",
       "name: \"LeakyRelu_197\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"700\"\n",
       "input: \"module_list.100.Conv2d.weight\"\n",
       "output: \"701\"\n",
       "name: \"Conv_198\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 224\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"701\"\n",
       "input: \"module_list.100.BatchNorm2d.weight\"\n",
       "input: \"module_list.100.BatchNorm2d.bias\"\n",
       "input: \"module_list.100.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.100.BatchNorm2d.running_var\"\n",
       "output: \"702\"\n",
       "name: \"BatchNormalization_199\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"702\"\n",
       "output: \"703\"\n",
       "name: \"LeakyRelu_200\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"703\"\n",
       "input: \"module_list.101.Conv2d.weight\"\n",
       "output: \"704\"\n",
       "name: \"Conv_201\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"704\"\n",
       "input: \"module_list.101.BatchNorm2d.weight\"\n",
       "input: \"module_list.101.BatchNorm2d.bias\"\n",
       "input: \"module_list.101.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.101.BatchNorm2d.running_var\"\n",
       "output: \"705\"\n",
       "name: \"BatchNormalization_202\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"705\"\n",
       "input: \"697\"\n",
       "output: \"706\"\n",
       "name: \"Add_203\"\n",
       "op_type: \"Add\"\n",
       ", input: \"706\"\n",
       "input: \"module_list.104.Conv2d.weight\"\n",
       "output: \"707\"\n",
       "name: \"Conv_204\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"707\"\n",
       "input: \"module_list.104.BatchNorm2d.weight\"\n",
       "input: \"module_list.104.BatchNorm2d.bias\"\n",
       "input: \"module_list.104.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.104.BatchNorm2d.running_var\"\n",
       "output: \"708\"\n",
       "name: \"BatchNormalization_205\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"708\"\n",
       "output: \"709\"\n",
       "name: \"LeakyRelu_206\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"709\"\n",
       "input: \"module_list.105.Conv2d.weight\"\n",
       "output: \"710\"\n",
       "name: \"Conv_207\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 224\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 3\n",
       "  ints: 3\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"710\"\n",
       "input: \"module_list.105.BatchNorm2d.weight\"\n",
       "input: \"module_list.105.BatchNorm2d.bias\"\n",
       "input: \"module_list.105.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.105.BatchNorm2d.running_var\"\n",
       "output: \"711\"\n",
       "name: \"BatchNormalization_208\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"711\"\n",
       "output: \"712\"\n",
       "name: \"LeakyRelu_209\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"712\"\n",
       "input: \"module_list.106.Conv2d.weight\"\n",
       "output: \"713\"\n",
       "name: \"Conv_210\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"713\"\n",
       "input: \"module_list.106.BatchNorm2d.weight\"\n",
       "input: \"module_list.106.BatchNorm2d.bias\"\n",
       "input: \"module_list.106.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.106.BatchNorm2d.running_var\"\n",
       "output: \"714\"\n",
       "name: \"BatchNormalization_211\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"714\"\n",
       "input: \"706\"\n",
       "output: \"715\"\n",
       "name: \"Add_212\"\n",
       "op_type: \"Add\"\n",
       ", input: \"715\"\n",
       "input: \"module_list.109.Conv2d.weight\"\n",
       "output: \"716\"\n",
       "name: \"Conv_213\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"716\"\n",
       "input: \"module_list.109.BatchNorm2d.weight\"\n",
       "input: \"module_list.109.BatchNorm2d.bias\"\n",
       "input: \"module_list.109.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.109.BatchNorm2d.running_var\"\n",
       "output: \"717\"\n",
       "name: \"BatchNormalization_214\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"717\"\n",
       "output: \"718\"\n",
       "name: \"LeakyRelu_215\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"718\"\n",
       "input: \"module_list.110.Conv2d.weight\"\n",
       "output: \"719\"\n",
       "name: \"Conv_216\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 5\n",
       "  ints: 5\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"719\"\n",
       "input: \"module_list.110.BatchNorm2d.weight\"\n",
       "input: \"module_list.110.BatchNorm2d.bias\"\n",
       "input: \"module_list.110.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.110.BatchNorm2d.running_var\"\n",
       "output: \"720\"\n",
       "name: \"BatchNormalization_217\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"720\"\n",
       "output: \"721\"\n",
       "name: \"LeakyRelu_218\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"721\"\n",
       "input: \"module_list.111.Conv2d.weight\"\n",
       "output: \"722\"\n",
       "name: \"Conv_219\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"722\"\n",
       "input: \"module_list.111.BatchNorm2d.weight\"\n",
       "input: \"module_list.111.BatchNorm2d.bias\"\n",
       "input: \"module_list.111.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.111.BatchNorm2d.running_var\"\n",
       "output: \"723\"\n",
       "name: \"BatchNormalization_220\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"723\"\n",
       "input: \"module_list.112.Conv2d.weight\"\n",
       "output: \"724\"\n",
       "name: \"Conv_221\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 128\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 5\n",
       "  ints: 5\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"724\"\n",
       "input: \"module_list.112.BatchNorm2d.weight\"\n",
       "input: \"module_list.112.BatchNorm2d.bias\"\n",
       "input: \"module_list.112.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.112.BatchNorm2d.running_var\"\n",
       "output: \"725\"\n",
       "name: \"BatchNormalization_222\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"725\"\n",
       "output: \"726\"\n",
       "name: \"LeakyRelu_223\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"726\"\n",
       "input: \"module_list.113.Conv2d.weight\"\n",
       "output: \"727\"\n",
       "name: \"Conv_224\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"727\"\n",
       "input: \"module_list.113.BatchNorm2d.weight\"\n",
       "input: \"module_list.113.BatchNorm2d.bias\"\n",
       "input: \"module_list.113.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.113.BatchNorm2d.running_var\"\n",
       "output: \"728\"\n",
       "name: \"BatchNormalization_225\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"728\"\n",
       "input: \"module_list.114.Conv2d.weight\"\n",
       "input: \"module_list.114.Conv2d.bias\"\n",
       "output: \"729\"\n",
       "name: \"Conv_226\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"729\"\n",
       "output: \"730\"\n",
       "name: \"Shape_227\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"731\"\n",
       "name: \"Constant_228\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"730\"\n",
       "input: \"731\"\n",
       "output: \"732\"\n",
       "name: \"Gather_229\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"732\"\n",
       "output: \"737\"\n",
       "name: \"Unsqueeze_230\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"737\"\n",
       "input: \"1536\"\n",
       "input: \"1537\"\n",
       "input: \"1538\"\n",
       "input: \"1539\"\n",
       "output: \"742\"\n",
       "name: \"Concat_231\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"729\"\n",
       "input: \"742\"\n",
       "output: \"743\"\n",
       "name: \"Reshape_232\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"743\"\n",
       "output: \"yolo1\"\n",
       "name: \"Transpose_233\"\n",
       "op_type: \"Transpose\"\n",
       "attribute {\n",
       "  name: \"perm\"\n",
       "  ints: 0\n",
       "  ints: 1\n",
       "  ints: 3\n",
       "  ints: 4\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       ", output: \"745\"\n",
       "name: \"Constant_234\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"746\"\n",
       "name: \"Constant_235\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"747\"\n",
       "name: \"Constant_236\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"748\"\n",
       "name: \"Constant_237\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"yolo1\"\n",
       "input: \"746\"\n",
       "input: \"747\"\n",
       "input: \"745\"\n",
       "input: \"748\"\n",
       "output: \"749\"\n",
       "name: \"Slice_238\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"749\"\n",
       "output: \"750\"\n",
       "name: \"Sigmoid_239\"\n",
       "op_type: \"Sigmoid\"\n",
       ", output: \"751\"\n",
       "name: \"Constant_240\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    dims: 1\n",
       "    dims: 10\n",
       "    dims: 16\n",
       "    dims: 2\n",
       "    data_type: 1\n",
       "    raw_data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`A\\000\\000\\000\\000\\000\\000pA\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\200?\\000\\000\\200?\\000\\000\\200?\\000\\000\\000@\\000\\000\\200?\\000\\000@@\\000\\000\\200?\\000\\000\\200@\\000\\000\\200?\\000\\000\\240@\\000\\000\\200?\\000\\000\\300@\\000\\000\\200?\\000\\000\\340@\\000\\000\\200?\\000\\000\\000A\\000\\000\\200?\\000\\000\\020A\\000\\000\\200?\\000\\000 A\\000\\000\\200?\\000\\0000A\\000\\000\\200?\\000\\000@A\\000\\000\\200?\\000\\000PA\\000\\000\\200?\\000\\000`A\\000\\000\\200?\\000\\000pA\\000\\000\\200?\\000\\000\\000\\000\\000\\000\\000@\\000\\000\\200?\\000\\000\\000@\\000\\000\\000@\\000\\000\\000@\\000\\000@@\\000\\000\\000@\\000\\000\\200@\\000\\000\\000@\\000\\000\\240@\\000\\000\\000@\\000\\000\\300@\\000\\000\\000@\\000\\000\\340@\\000\\000\\000@\\000\\000\\000A\\000\\000\\000@\\000\\000\\020A\\000\\000\\000@\\000\\000 A\\000\\000\\000@\\000\\0000A\\000\\000\\000@\\000\\000@A\\000\\000\\000@\\000\\000PA\\000\\000\\000@\\000\\000`A\\000\\000\\000@\\000\\000pA\\000\\000\\000@\\000\\000\\000\\000\\000\\000@@\\000\\000\\200?\\000\\000@@\\000\\000\\000@\\000\\000@@\\000\\000@@\\000\\000@@\\000\\000\\200@\\000\\000@@\\000\\000\\240@\\000\\000@@\\000\\000\\300@\\000\\000@@\\000\\000\\340@\\000\\000@@\\000\\000\\000A\\000\\000@@\\000\\000\\020A\\000\\000@@\\000\\000 A\\000\\000@@\\000\\0000A\\000\\000@@\\000\\000@A\\000\\000@@\\000\\000PA\\000\\000@@\\000\\000`A\\000\\000@@\\000\\000pA\\000\\000@@\\000\\000\\000\\000\\000\\000\\200@\\000\\000\\200?\\000\\000\\200@\\000\\000\\000@\\000\\000\\200@\\000\\000@@\\000\\000\\200@\\000\\000\\200@\\000\\000\\200@\\000\\000\\240@\\000\\000\\200@\\000\\000\\300@\\000\\000\\200@\\000\\000\\340@\\000\\000\\200@\\000\\000\\000A\\000\\000\\200@\\000\\000\\020A\\000\\000\\200@\\000\\000 A\\000\\000\\200@\\000\\0000A\\000\\000\\200@\\000\\000@A\\000\\000\\200@\\000\\000PA\\000\\000\\200@\\000\\000`A\\000\\000\\200@\\000\\000pA\\000\\000\\200@\\000\\000\\000\\000\\000\\000\\240@\\000\\000\\200?\\000\\000\\240@\\000\\000\\000@\\000\\000\\240@\\000\\000@@\\000\\000\\240@\\000\\000\\200@\\000\\000\\240@\\000\\000\\240@\\000\\000\\240@\\000\\000\\300@\\000\\000\\240@\\000\\000\\340@\\000\\000\\240@\\000\\000\\000A\\000\\000\\240@\\000\\000\\020A\\000\\000\\240@\\000\\000 A\\000\\000\\240@\\000\\0000A\\000\\000\\240@\\000\\000@A\\000\\000\\240@\\000\\000PA\\000\\000\\240@\\000\\000`A\\000\\000\\240@\\000\\000pA\\000\\000\\240@\\000\\000\\000\\000\\000\\000\\300@\\000\\000\\200?\\000\\000\\300@\\000\\000\\000@\\000\\000\\300@\\000\\000@@\\000\\000\\300@\\000\\000\\200@\\000\\000\\300@\\000\\000\\240@\\000\\000\\300@\\000\\000\\300@\\000\\000\\300@\\000\\000\\340@\\000\\000\\300@\\000\\000\\000A\\000\\000\\300@\\000\\000\\020A\\000\\000\\300@\\000\\000 A\\000\\000\\300@\\000\\0000A\\000\\000\\300@\\000\\000@A\\000\\000\\300@\\000\\000PA\\000\\000\\300@\\000\\000`A\\000\\000\\300@\\000\\000pA\\000\\000\\300@\\000\\000\\000\\000\\000\\000\\340@\\000\\000\\200?\\000\\000\\340@\\000\\000\\000@\\000\\000\\340@\\000\\000@@\\000\\000\\340@\\000\\000\\200@\\000\\000\\340@\\000\\000\\240@\\000\\000\\340@\\000\\000\\300@\\000\\000\\340@\\000\\000\\340@\\000\\000\\340@\\000\\000\\000A\\000\\000\\340@\\000\\000\\020A\\000\\000\\340@\\000\\000 A\\000\\000\\340@\\000\\0000A\\000\\000\\340@\\000\\000@A\\000\\000\\340@\\000\\000PA\\000\\000\\340@\\000\\000`A\\000\\000\\340@\\000\\000pA\\000\\000\\340@\\000\\000\\000\\000\\000\\000\\000A\\000\\000\\200?\\000\\000\\000A\\000\\000\\000@\\000\\000\\000A\\000\\000@@\\000\\000\\000A\\000\\000\\200@\\000\\000\\000A\\000\\000\\240@\\000\\000\\000A\\000\\000\\300@\\000\\000\\000A\\000\\000\\340@\\000\\000\\000A\\000\\000\\000A\\000\\000\\000A\\000\\000\\020A\\000\\000\\000A\\000\\000 A\\000\\000\\000A\\000\\0000A\\000\\000\\000A\\000\\000@A\\000\\000\\000A\\000\\000PA\\000\\000\\000A\\000\\000`A\\000\\000\\000A\\000\\000pA\\000\\000\\000A\\000\\000\\000\\000\\000\\000\\020A\\000\\000\\200?\\000\\000\\020A\\000\\000\\000@\\000\\000\\020A\\000\\000@@\\000\\000\\020A\\000\\000\\200@\\000\\000\\020A\\000\\000\\240@\\000\\000\\020A\\000\\000\\300@\\000\\000\\020A\\000\\000\\340@\\000\\000\\020A\\000\\000\\000A\\000\\000\\020A\\000\\000\\020A\\000\\000\\020A\\000\\000 A\\000\\000\\020A\\000\\0000A\\000\\000\\020A\\000\\000@A\\000\\000\\020A\\000\\000PA\\000\\000\\020A\\000\\000`A\\000\\000\\020A\\000\\000pA\\000\\000\\020A\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"750\"\n",
       "input: \"751\"\n",
       "output: \"752\"\n",
       "name: \"Add_241\"\n",
       "op_type: \"Add\"\n",
       ", output: \"753\"\n",
       "name: \"Constant_242\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"752\"\n",
       "input: \"753\"\n",
       "output: \"754\"\n",
       "name: \"Reshape_243\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"755\"\n",
       "name: \"Constant_244\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1540\"\n",
       "output: \"757\"\n",
       "name: \"ConstantOfShape_245\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"758\"\n",
       "name: \"Constant_246\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"757\"\n",
       "input: \"758\"\n",
       "output: \"759\"\n",
       "name: \"Mul_247\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"760\"\n",
       "name: \"Constant_248\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"760\"\n",
       "input: \"759\"\n",
       "output: \"761\"\n",
       "name: \"Equal_249\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"761\"\n",
       "input: \"757\"\n",
       "input: \"755\"\n",
       "output: \"762\"\n",
       "name: \"Where_250\"\n",
       "op_type: \"Where\"\n",
       ", input: \"754\"\n",
       "input: \"762\"\n",
       "output: \"763\"\n",
       "name: \"Expand_251\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"yolo1\"\n",
       "output: \"764\"\n",
       "name: \"Shape_252\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"765\"\n",
       "name: \"Constant_253\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"764\"\n",
       "input: \"765\"\n",
       "output: \"766\"\n",
       "name: \"Gather_254\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"766\"\n",
       "output: \"767\"\n",
       "name: \"Cast_255\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"768\"\n",
       "name: \"Constant_256\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"769\"\n",
       "name: \"Constant_257\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"768\"\n",
       "input: \"767\"\n",
       "input: \"769\"\n",
       "output: \"770\"\n",
       "name: \"Range_258\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo1\"\n",
       "output: \"771\"\n",
       "name: \"Shape_259\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"772\"\n",
       "name: \"Constant_260\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"771\"\n",
       "input: \"772\"\n",
       "output: \"773\"\n",
       "name: \"Gather_261\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"773\"\n",
       "output: \"774\"\n",
       "name: \"Cast_262\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"775\"\n",
       "name: \"Constant_263\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"776\"\n",
       "name: \"Constant_264\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"775\"\n",
       "input: \"774\"\n",
       "input: \"776\"\n",
       "output: \"777\"\n",
       "name: \"Range_265\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo1\"\n",
       "output: \"778\"\n",
       "name: \"Shape_266\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"779\"\n",
       "name: \"Constant_267\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"778\"\n",
       "input: \"779\"\n",
       "output: \"780\"\n",
       "name: \"Gather_268\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"780\"\n",
       "output: \"781\"\n",
       "name: \"Cast_269\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"782\"\n",
       "name: \"Constant_270\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"783\"\n",
       "name: \"Constant_271\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"782\"\n",
       "input: \"781\"\n",
       "input: \"783\"\n",
       "output: \"784\"\n",
       "name: \"Range_272\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo1\"\n",
       "output: \"785\"\n",
       "name: \"Shape_273\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"786\"\n",
       "name: \"Constant_274\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"785\"\n",
       "input: \"786\"\n",
       "output: \"787\"\n",
       "name: \"Gather_275\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"787\"\n",
       "output: \"788\"\n",
       "name: \"Cast_276\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"789\"\n",
       "name: \"Constant_277\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"790\"\n",
       "name: \"Constant_278\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"789\"\n",
       "input: \"788\"\n",
       "input: \"790\"\n",
       "output: \"791\"\n",
       "name: \"Range_279\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo1\"\n",
       "output: \"792\"\n",
       "name: \"Shape_280\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"793\"\n",
       "name: \"Constant_281\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"792\"\n",
       "input: \"793\"\n",
       "output: \"794\"\n",
       "name: \"Gather_282\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"794\"\n",
       "output: \"795\"\n",
       "name: \"Cast_283\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"796\"\n",
       "name: \"Constant_284\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"797\"\n",
       "name: \"Constant_285\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"796\"\n",
       "input: \"795\"\n",
       "input: \"797\"\n",
       "output: \"798\"\n",
       "name: \"Range_286\"\n",
       "op_type: \"Range\"\n",
       ", output: \"799\"\n",
       "name: \"Constant_287\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"800\"\n",
       "name: \"Constant_288\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"801\"\n",
       "name: \"Constant_289\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"802\"\n",
       "name: \"Constant_290\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"798\"\n",
       "input: \"800\"\n",
       "input: \"801\"\n",
       "input: \"799\"\n",
       "input: \"802\"\n",
       "output: \"803\"\n",
       "name: \"Slice_291\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"804\"\n",
       "name: \"Constant_292\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"770\"\n",
       "input: \"804\"\n",
       "output: \"805\"\n",
       "name: \"Reshape_293\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"806\"\n",
       "name: \"Constant_294\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"777\"\n",
       "input: \"806\"\n",
       "output: \"807\"\n",
       "name: \"Reshape_295\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"808\"\n",
       "name: \"Constant_296\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 3\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"784\"\n",
       "input: \"808\"\n",
       "output: \"809\"\n",
       "name: \"Reshape_297\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"810\"\n",
       "name: \"Constant_298\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 2\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"791\"\n",
       "input: \"810\"\n",
       "output: \"811\"\n",
       "name: \"Reshape_299\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"812\"\n",
       "name: \"Constant_300\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"803\"\n",
       "input: \"812\"\n",
       "output: \"813\"\n",
       "name: \"Reshape_301\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"805\"\n",
       "input: \"807\"\n",
       "output: \"814\"\n",
       "name: \"Add_302\"\n",
       "op_type: \"Add\"\n",
       ", input: \"814\"\n",
       "input: \"809\"\n",
       "output: \"815\"\n",
       "name: \"Add_303\"\n",
       "op_type: \"Add\"\n",
       ", input: \"815\"\n",
       "input: \"811\"\n",
       "output: \"816\"\n",
       "name: \"Add_304\"\n",
       "op_type: \"Add\"\n",
       ", input: \"816\"\n",
       "input: \"813\"\n",
       "output: \"817\"\n",
       "name: \"Add_305\"\n",
       "op_type: \"Add\"\n",
       ", input: \"817\"\n",
       "output: \"818\"\n",
       "name: \"Shape_306\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"818\"\n",
       "output: \"819\"\n",
       "name: \"Shape_307\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"819\"\n",
       "output: \"820\"\n",
       "name: \"ConstantOfShape_308\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"821\"\n",
       "name: \"Constant_309\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"820\"\n",
       "input: \"821\"\n",
       "output: \"822\"\n",
       "name: \"Mul_310\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"818\"\n",
       "input: \"822\"\n",
       "output: \"823\"\n",
       "name: \"Equal_311\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"823\"\n",
       "input: \"820\"\n",
       "input: \"818\"\n",
       "output: \"824\"\n",
       "name: \"Where_312\"\n",
       "op_type: \"Where\"\n",
       ", input: \"805\"\n",
       "input: \"824\"\n",
       "output: \"825\"\n",
       "name: \"Expand_313\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"825\"\n",
       "output: \"826\"\n",
       "name: \"Unsqueeze_314\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"818\"\n",
       "output: \"827\"\n",
       "name: \"Shape_315\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"827\"\n",
       "output: \"828\"\n",
       "name: \"ConstantOfShape_316\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"829\"\n",
       "name: \"Constant_317\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"828\"\n",
       "input: \"829\"\n",
       "output: \"830\"\n",
       "name: \"Mul_318\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"818\"\n",
       "input: \"830\"\n",
       "output: \"831\"\n",
       "name: \"Equal_319\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"831\"\n",
       "input: \"828\"\n",
       "input: \"818\"\n",
       "output: \"832\"\n",
       "name: \"Where_320\"\n",
       "op_type: \"Where\"\n",
       ", input: \"807\"\n",
       "input: \"832\"\n",
       "output: \"833\"\n",
       "name: \"Expand_321\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"833\"\n",
       "output: \"834\"\n",
       "name: \"Unsqueeze_322\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"818\"\n",
       "output: \"835\"\n",
       "name: \"Shape_323\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"835\"\n",
       "output: \"836\"\n",
       "name: \"ConstantOfShape_324\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"837\"\n",
       "name: \"Constant_325\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"836\"\n",
       "input: \"837\"\n",
       "output: \"838\"\n",
       "name: \"Mul_326\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"818\"\n",
       "input: \"838\"\n",
       "output: \"839\"\n",
       "name: \"Equal_327\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"839\"\n",
       "input: \"836\"\n",
       "input: \"818\"\n",
       "output: \"840\"\n",
       "name: \"Where_328\"\n",
       "op_type: \"Where\"\n",
       ", input: \"809\"\n",
       "input: \"840\"\n",
       "output: \"841\"\n",
       "name: \"Expand_329\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"841\"\n",
       "output: \"842\"\n",
       "name: \"Unsqueeze_330\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"818\"\n",
       "output: \"843\"\n",
       "name: \"Shape_331\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"843\"\n",
       "output: \"844\"\n",
       "name: \"ConstantOfShape_332\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"845\"\n",
       "name: \"Constant_333\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"844\"\n",
       "input: \"845\"\n",
       "output: \"846\"\n",
       "name: \"Mul_334\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"818\"\n",
       "input: \"846\"\n",
       "output: \"847\"\n",
       "name: \"Equal_335\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"847\"\n",
       "input: \"844\"\n",
       "input: \"818\"\n",
       "output: \"848\"\n",
       "name: \"Where_336\"\n",
       "op_type: \"Where\"\n",
       ", input: \"811\"\n",
       "input: \"848\"\n",
       "output: \"849\"\n",
       "name: \"Expand_337\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"849\"\n",
       "output: \"850\"\n",
       "name: \"Unsqueeze_338\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"818\"\n",
       "output: \"851\"\n",
       "name: \"Shape_339\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"851\"\n",
       "output: \"852\"\n",
       "name: \"ConstantOfShape_340\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"853\"\n",
       "name: \"Constant_341\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"852\"\n",
       "input: \"853\"\n",
       "output: \"854\"\n",
       "name: \"Mul_342\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"818\"\n",
       "input: \"854\"\n",
       "output: \"855\"\n",
       "name: \"Equal_343\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"855\"\n",
       "input: \"852\"\n",
       "input: \"818\"\n",
       "output: \"856\"\n",
       "name: \"Where_344\"\n",
       "op_type: \"Where\"\n",
       ", input: \"813\"\n",
       "input: \"856\"\n",
       "output: \"857\"\n",
       "name: \"Expand_345\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"857\"\n",
       "output: \"858\"\n",
       "name: \"Unsqueeze_346\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"826\"\n",
       "input: \"834\"\n",
       "input: \"842\"\n",
       "input: \"850\"\n",
       "input: \"858\"\n",
       "output: \"859\"\n",
       "name: \"Concat_347\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: -1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"yolo1\"\n",
       "output: \"860\"\n",
       "name: \"Shape_348\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"861\"\n",
       "name: \"Constant_349\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"862\"\n",
       "name: \"Constant_350\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\005\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"863\"\n",
       "name: \"Constant_351\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\177\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"860\"\n",
       "input: \"862\"\n",
       "input: \"863\"\n",
       "input: \"861\"\n",
       "output: \"864\"\n",
       "name: \"Slice_352\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"818\"\n",
       "input: \"864\"\n",
       "output: \"865\"\n",
       "name: \"Concat_353\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"763\"\n",
       "input: \"865\"\n",
       "output: \"866\"\n",
       "name: \"Reshape_354\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"yolo1\"\n",
       "input: \"859\"\n",
       "input: \"866\"\n",
       "output: \"867\"\n",
       "name: \"ScatterND_355\"\n",
       "op_type: \"ScatterND\"\n",
       ", output: \"868\"\n",
       "name: \"Constant_356\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"869\"\n",
       "name: \"Constant_357\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"870\"\n",
       "name: \"Constant_358\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"871\"\n",
       "name: \"Constant_359\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"867\"\n",
       "input: \"869\"\n",
       "input: \"870\"\n",
       "input: \"868\"\n",
       "input: \"871\"\n",
       "output: \"872\"\n",
       "name: \"Slice_360\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"872\"\n",
       "output: \"873\"\n",
       "name: \"Exp_361\"\n",
       "op_type: \"Exp\"\n",
       ", output: \"874\"\n",
       "name: \"Constant_362\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    dims: 3\n",
       "    dims: 1\n",
       "    dims: 1\n",
       "    dims: 2\n",
       "    data_type: 1\n",
       "    raw_data: \"\\000\\000f@\\000\\000\\022@\\000\\000n@\\000\\000\\307@\\000\\000\\362@\\000\\000\\356@\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"873\"\n",
       "input: \"874\"\n",
       "output: \"875\"\n",
       "name: \"Mul_363\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"876\"\n",
       "name: \"Constant_364\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"875\"\n",
       "input: \"876\"\n",
       "output: \"877\"\n",
       "name: \"Reshape_365\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"878\"\n",
       "name: \"Constant_366\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1541\"\n",
       "output: \"880\"\n",
       "name: \"ConstantOfShape_367\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"881\"\n",
       "name: \"Constant_368\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"880\"\n",
       "input: \"881\"\n",
       "output: \"882\"\n",
       "name: \"Mul_369\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"883\"\n",
       "name: \"Constant_370\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"883\"\n",
       "input: \"882\"\n",
       "output: \"884\"\n",
       "name: \"Equal_371\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"884\"\n",
       "input: \"880\"\n",
       "input: \"878\"\n",
       "output: \"885\"\n",
       "name: \"Where_372\"\n",
       "op_type: \"Where\"\n",
       ", input: \"877\"\n",
       "input: \"885\"\n",
       "output: \"886\"\n",
       "name: \"Expand_373\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"867\"\n",
       "output: \"887\"\n",
       "name: \"Shape_374\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"888\"\n",
       "name: \"Constant_375\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"887\"\n",
       "input: \"888\"\n",
       "output: \"889\"\n",
       "name: \"Gather_376\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"889\"\n",
       "output: \"890\"\n",
       "name: \"Cast_377\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"891\"\n",
       "name: \"Constant_378\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"892\"\n",
       "name: \"Constant_379\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"891\"\n",
       "input: \"890\"\n",
       "input: \"892\"\n",
       "output: \"893\"\n",
       "name: \"Range_380\"\n",
       "op_type: \"Range\"\n",
       ", input: \"867\"\n",
       "output: \"894\"\n",
       "name: \"Shape_381\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"895\"\n",
       "name: \"Constant_382\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"894\"\n",
       "input: \"895\"\n",
       "output: \"896\"\n",
       "name: \"Gather_383\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"896\"\n",
       "output: \"897\"\n",
       "name: \"Cast_384\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"898\"\n",
       "name: \"Constant_385\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"899\"\n",
       "name: \"Constant_386\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"898\"\n",
       "input: \"897\"\n",
       "input: \"899\"\n",
       "output: \"900\"\n",
       "name: \"Range_387\"\n",
       "op_type: \"Range\"\n",
       ", input: \"867\"\n",
       "output: \"901\"\n",
       "name: \"Shape_388\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"902\"\n",
       "name: \"Constant_389\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"901\"\n",
       "input: \"902\"\n",
       "output: \"903\"\n",
       "name: \"Gather_390\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"903\"\n",
       "output: \"904\"\n",
       "name: \"Cast_391\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"905\"\n",
       "name: \"Constant_392\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"906\"\n",
       "name: \"Constant_393\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"905\"\n",
       "input: \"904\"\n",
       "input: \"906\"\n",
       "output: \"907\"\n",
       "name: \"Range_394\"\n",
       "op_type: \"Range\"\n",
       ", input: \"867\"\n",
       "output: \"908\"\n",
       "name: \"Shape_395\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"909\"\n",
       "name: \"Constant_396\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"908\"\n",
       "input: \"909\"\n",
       "output: \"910\"\n",
       "name: \"Gather_397\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"910\"\n",
       "output: \"911\"\n",
       "name: \"Cast_398\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"912\"\n",
       "name: \"Constant_399\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"913\"\n",
       "name: \"Constant_400\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"912\"\n",
       "input: \"911\"\n",
       "input: \"913\"\n",
       "output: \"914\"\n",
       "name: \"Range_401\"\n",
       "op_type: \"Range\"\n",
       ", input: \"867\"\n",
       "output: \"915\"\n",
       "name: \"Shape_402\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"916\"\n",
       "name: \"Constant_403\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"915\"\n",
       "input: \"916\"\n",
       "output: \"917\"\n",
       "name: \"Gather_404\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"917\"\n",
       "output: \"918\"\n",
       "name: \"Cast_405\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"919\"\n",
       "name: \"Constant_406\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"920\"\n",
       "name: \"Constant_407\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"919\"\n",
       "input: \"918\"\n",
       "input: \"920\"\n",
       "output: \"921\"\n",
       "name: \"Range_408\"\n",
       "op_type: \"Range\"\n",
       ", output: \"922\"\n",
       "name: \"Constant_409\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"923\"\n",
       "name: \"Constant_410\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"924\"\n",
       "name: \"Constant_411\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"925\"\n",
       "name: \"Constant_412\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"921\"\n",
       "input: \"923\"\n",
       "input: \"924\"\n",
       "input: \"922\"\n",
       "input: \"925\"\n",
       "output: \"926\"\n",
       "name: \"Slice_413\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"927\"\n",
       "name: \"Constant_414\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"893\"\n",
       "input: \"927\"\n",
       "output: \"928\"\n",
       "name: \"Reshape_415\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"929\"\n",
       "name: \"Constant_416\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"900\"\n",
       "input: \"929\"\n",
       "output: \"930\"\n",
       "name: \"Reshape_417\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"931\"\n",
       "name: \"Constant_418\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 3\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"907\"\n",
       "input: \"931\"\n",
       "output: \"932\"\n",
       "name: \"Reshape_419\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"933\"\n",
       "name: \"Constant_420\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 2\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"914\"\n",
       "input: \"933\"\n",
       "output: \"934\"\n",
       "name: \"Reshape_421\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"935\"\n",
       "name: \"Constant_422\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"926\"\n",
       "input: \"935\"\n",
       "output: \"936\"\n",
       "name: \"Reshape_423\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"928\"\n",
       "input: \"930\"\n",
       "output: \"937\"\n",
       "name: \"Add_424\"\n",
       "op_type: \"Add\"\n",
       ", input: \"937\"\n",
       "input: \"932\"\n",
       "output: \"938\"\n",
       "name: \"Add_425\"\n",
       "op_type: \"Add\"\n",
       ", input: \"938\"\n",
       "input: \"934\"\n",
       "output: \"939\"\n",
       "name: \"Add_426\"\n",
       "op_type: \"Add\"\n",
       ", input: \"939\"\n",
       "input: \"936\"\n",
       "output: \"940\"\n",
       "name: \"Add_427\"\n",
       "op_type: \"Add\"\n",
       ", input: \"940\"\n",
       "output: \"941\"\n",
       "name: \"Shape_428\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"941\"\n",
       "output: \"942\"\n",
       "name: \"Shape_429\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"942\"\n",
       "output: \"943\"\n",
       "name: \"ConstantOfShape_430\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"944\"\n",
       "name: \"Constant_431\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"943\"\n",
       "input: \"944\"\n",
       "output: \"945\"\n",
       "name: \"Mul_432\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"941\"\n",
       "input: \"945\"\n",
       "output: \"946\"\n",
       "name: \"Equal_433\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"946\"\n",
       "input: \"943\"\n",
       "input: \"941\"\n",
       "output: \"947\"\n",
       "name: \"Where_434\"\n",
       "op_type: \"Where\"\n",
       ", input: \"928\"\n",
       "input: \"947\"\n",
       "output: \"948\"\n",
       "name: \"Expand_435\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"948\"\n",
       "output: \"949\"\n",
       "name: \"Unsqueeze_436\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"941\"\n",
       "output: \"950\"\n",
       "name: \"Shape_437\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"950\"\n",
       "output: \"951\"\n",
       "name: \"ConstantOfShape_438\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"952\"\n",
       "name: \"Constant_439\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"951\"\n",
       "input: \"952\"\n",
       "output: \"953\"\n",
       "name: \"Mul_440\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"941\"\n",
       "input: \"953\"\n",
       "output: \"954\"\n",
       "name: \"Equal_441\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"954\"\n",
       "input: \"951\"\n",
       "input: \"941\"\n",
       "output: \"955\"\n",
       "name: \"Where_442\"\n",
       "op_type: \"Where\"\n",
       ", input: \"930\"\n",
       "input: \"955\"\n",
       "output: \"956\"\n",
       "name: \"Expand_443\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"956\"\n",
       "output: \"957\"\n",
       "name: \"Unsqueeze_444\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"941\"\n",
       "output: \"958\"\n",
       "name: \"Shape_445\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"958\"\n",
       "output: \"959\"\n",
       "name: \"ConstantOfShape_446\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"960\"\n",
       "name: \"Constant_447\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"959\"\n",
       "input: \"960\"\n",
       "output: \"961\"\n",
       "name: \"Mul_448\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"941\"\n",
       "input: \"961\"\n",
       "output: \"962\"\n",
       "name: \"Equal_449\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"962\"\n",
       "input: \"959\"\n",
       "input: \"941\"\n",
       "output: \"963\"\n",
       "name: \"Where_450\"\n",
       "op_type: \"Where\"\n",
       ", input: \"932\"\n",
       "input: \"963\"\n",
       "output: \"964\"\n",
       "name: \"Expand_451\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"964\"\n",
       "output: \"965\"\n",
       "name: \"Unsqueeze_452\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"941\"\n",
       "output: \"966\"\n",
       "name: \"Shape_453\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"966\"\n",
       "output: \"967\"\n",
       "name: \"ConstantOfShape_454\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"968\"\n",
       "name: \"Constant_455\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"967\"\n",
       "input: \"968\"\n",
       "output: \"969\"\n",
       "name: \"Mul_456\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"941\"\n",
       "input: \"969\"\n",
       "output: \"970\"\n",
       "name: \"Equal_457\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"970\"\n",
       "input: \"967\"\n",
       "input: \"941\"\n",
       "output: \"971\"\n",
       "name: \"Where_458\"\n",
       "op_type: \"Where\"\n",
       ", input: \"934\"\n",
       "input: \"971\"\n",
       "output: \"972\"\n",
       "name: \"Expand_459\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"972\"\n",
       "output: \"973\"\n",
       "name: \"Unsqueeze_460\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"941\"\n",
       "output: \"974\"\n",
       "name: \"Shape_461\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"974\"\n",
       "output: \"975\"\n",
       "name: \"ConstantOfShape_462\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"976\"\n",
       "name: \"Constant_463\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"975\"\n",
       "input: \"976\"\n",
       "output: \"977\"\n",
       "name: \"Mul_464\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"941\"\n",
       "input: \"977\"\n",
       "output: \"978\"\n",
       "name: \"Equal_465\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"978\"\n",
       "input: \"975\"\n",
       "input: \"941\"\n",
       "output: \"979\"\n",
       "name: \"Where_466\"\n",
       "op_type: \"Where\"\n",
       ", input: \"936\"\n",
       "input: \"979\"\n",
       "output: \"980\"\n",
       "name: \"Expand_467\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"980\"\n",
       "output: \"981\"\n",
       "name: \"Unsqueeze_468\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"949\"\n",
       "input: \"957\"\n",
       "input: \"965\"\n",
       "input: \"973\"\n",
       "input: \"981\"\n",
       "output: \"982\"\n",
       "name: \"Concat_469\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: -1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"867\"\n",
       "output: \"983\"\n",
       "name: \"Shape_470\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"984\"\n",
       "name: \"Constant_471\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"985\"\n",
       "name: \"Constant_472\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\005\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"986\"\n",
       "name: \"Constant_473\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\177\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"983\"\n",
       "input: \"985\"\n",
       "input: \"986\"\n",
       "input: \"984\"\n",
       "output: \"987\"\n",
       "name: \"Slice_474\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"941\"\n",
       "input: \"987\"\n",
       "output: \"988\"\n",
       "name: \"Concat_475\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"886\"\n",
       "input: \"988\"\n",
       "output: \"989\"\n",
       "name: \"Reshape_476\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"867\"\n",
       "input: \"982\"\n",
       "input: \"989\"\n",
       "output: \"990\"\n",
       "name: \"ScatterND_477\"\n",
       "op_type: \"ScatterND\"\n",
       ", output: \"991\"\n",
       "name: \"Constant_478\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"992\"\n",
       "name: \"Constant_479\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"993\"\n",
       "name: \"Constant_480\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"994\"\n",
       "name: \"Constant_481\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"990\"\n",
       "input: \"992\"\n",
       "input: \"993\"\n",
       "input: \"991\"\n",
       "input: \"994\"\n",
       "output: \"995\"\n",
       "name: \"Slice_482\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"996\"\n",
       "name: \"Constant_483\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 1\n",
       "    raw_data: \"\\000\\000\\000B\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"995\"\n",
       "input: \"996\"\n",
       "output: \"997\"\n",
       "name: \"Mul_484\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"998\"\n",
       "name: \"Constant_485\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"997\"\n",
       "input: \"998\"\n",
       "output: \"999\"\n",
       "name: \"Reshape_486\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1000\"\n",
       "name: \"Constant_487\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1542\"\n",
       "output: \"1002\"\n",
       "name: \"ConstantOfShape_488\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1003\"\n",
       "name: \"Constant_489\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1002\"\n",
       "input: \"1003\"\n",
       "output: \"1004\"\n",
       "name: \"Mul_490\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"1005\"\n",
       "name: \"Constant_491\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\n\\000\\000\\000\\000\\000\\000\\000\\020\\000\\000\\000\\000\\000\\000\\000\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1005\"\n",
       "input: \"1004\"\n",
       "output: \"1006\"\n",
       "name: \"Equal_492\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1006\"\n",
       "input: \"1002\"\n",
       "input: \"1000\"\n",
       "output: \"1007\"\n",
       "name: \"Where_493\"\n",
       "op_type: \"Where\"\n",
       ", input: \"999\"\n",
       "input: \"1007\"\n",
       "output: \"1008\"\n",
       "name: \"Expand_494\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"990\"\n",
       "output: \"1009\"\n",
       "name: \"Shape_495\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1010\"\n",
       "name: \"Constant_496\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1009\"\n",
       "input: \"1010\"\n",
       "output: \"1011\"\n",
       "name: \"Gather_497\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1011\"\n",
       "output: \"1012\"\n",
       "name: \"Cast_498\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1013\"\n",
       "name: \"Constant_499\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1014\"\n",
       "name: \"Constant_500\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1013\"\n",
       "input: \"1012\"\n",
       "input: \"1014\"\n",
       "output: \"1015\"\n",
       "name: \"Range_501\"\n",
       "op_type: \"Range\"\n",
       ", input: \"990\"\n",
       "output: \"1016\"\n",
       "name: \"Shape_502\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1017\"\n",
       "name: \"Constant_503\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1016\"\n",
       "input: \"1017\"\n",
       "output: \"1018\"\n",
       "name: \"Gather_504\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1018\"\n",
       "output: \"1019\"\n",
       "name: \"Cast_505\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1020\"\n",
       "name: \"Constant_506\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1021\"\n",
       "name: \"Constant_507\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1020\"\n",
       "input: \"1019\"\n",
       "input: \"1021\"\n",
       "output: \"1022\"\n",
       "name: \"Range_508\"\n",
       "op_type: \"Range\"\n",
       ", input: \"990\"\n",
       "output: \"1023\"\n",
       "name: \"Shape_509\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1024\"\n",
       "name: \"Constant_510\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1023\"\n",
       "input: \"1024\"\n",
       "output: \"1025\"\n",
       "name: \"Gather_511\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1025\"\n",
       "output: \"1026\"\n",
       "name: \"Cast_512\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1027\"\n",
       "name: \"Constant_513\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1028\"\n",
       "name: \"Constant_514\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1027\"\n",
       "input: \"1026\"\n",
       "input: \"1028\"\n",
       "output: \"1029\"\n",
       "name: \"Range_515\"\n",
       "op_type: \"Range\"\n",
       ", input: \"990\"\n",
       "output: \"1030\"\n",
       "name: \"Shape_516\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1031\"\n",
       "name: \"Constant_517\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1030\"\n",
       "input: \"1031\"\n",
       "output: \"1032\"\n",
       "name: \"Gather_518\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1032\"\n",
       "output: \"1033\"\n",
       "name: \"Cast_519\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1034\"\n",
       "name: \"Constant_520\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1035\"\n",
       "name: \"Constant_521\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1034\"\n",
       "input: \"1033\"\n",
       "input: \"1035\"\n",
       "output: \"1036\"\n",
       "name: \"Range_522\"\n",
       "op_type: \"Range\"\n",
       ", input: \"990\"\n",
       "output: \"1037\"\n",
       "name: \"Shape_523\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1038\"\n",
       "name: \"Constant_524\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1037\"\n",
       "input: \"1038\"\n",
       "output: \"1039\"\n",
       "name: \"Gather_525\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1039\"\n",
       "output: \"1040\"\n",
       "name: \"Cast_526\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1041\"\n",
       "name: \"Constant_527\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1042\"\n",
       "name: \"Constant_528\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1041\"\n",
       "input: \"1040\"\n",
       "input: \"1042\"\n",
       "output: \"1043\"\n",
       "name: \"Range_529\"\n",
       "op_type: \"Range\"\n",
       ", output: \"1044\"\n",
       "name: \"Constant_530\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1045\"\n",
       "name: \"Constant_531\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1046\"\n",
       "name: \"Constant_532\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1047\"\n",
       "name: \"Constant_533\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1043\"\n",
       "input: \"1045\"\n",
       "input: \"1046\"\n",
       "input: \"1044\"\n",
       "input: \"1047\"\n",
       "output: \"1048\"\n",
       "name: \"Slice_534\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"1049\"\n",
       "name: \"Constant_535\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1015\"\n",
       "input: \"1049\"\n",
       "output: \"1050\"\n",
       "name: \"Reshape_536\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1051\"\n",
       "name: \"Constant_537\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1022\"\n",
       "input: \"1051\"\n",
       "output: \"1052\"\n",
       "name: \"Reshape_538\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1053\"\n",
       "name: \"Constant_539\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 3\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1029\"\n",
       "input: \"1053\"\n",
       "output: \"1054\"\n",
       "name: \"Reshape_540\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1055\"\n",
       "name: \"Constant_541\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 2\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1036\"\n",
       "input: \"1055\"\n",
       "output: \"1056\"\n",
       "name: \"Reshape_542\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1057\"\n",
       "name: \"Constant_543\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1048\"\n",
       "input: \"1057\"\n",
       "output: \"1058\"\n",
       "name: \"Reshape_544\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1050\"\n",
       "input: \"1052\"\n",
       "output: \"1059\"\n",
       "name: \"Add_545\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1059\"\n",
       "input: \"1054\"\n",
       "output: \"1060\"\n",
       "name: \"Add_546\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1060\"\n",
       "input: \"1056\"\n",
       "output: \"1061\"\n",
       "name: \"Add_547\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1061\"\n",
       "input: \"1058\"\n",
       "output: \"1062\"\n",
       "name: \"Add_548\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1062\"\n",
       "output: \"1063\"\n",
       "name: \"Shape_549\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1063\"\n",
       "output: \"1064\"\n",
       "name: \"Shape_550\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1064\"\n",
       "output: \"1065\"\n",
       "name: \"ConstantOfShape_551\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1066\"\n",
       "name: \"Constant_552\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1065\"\n",
       "input: \"1066\"\n",
       "output: \"1067\"\n",
       "name: \"Mul_553\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1063\"\n",
       "input: \"1067\"\n",
       "output: \"1068\"\n",
       "name: \"Equal_554\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1068\"\n",
       "input: \"1065\"\n",
       "input: \"1063\"\n",
       "output: \"1069\"\n",
       "name: \"Where_555\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1050\"\n",
       "input: \"1069\"\n",
       "output: \"1070\"\n",
       "name: \"Expand_556\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1070\"\n",
       "output: \"1071\"\n",
       "name: \"Unsqueeze_557\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1063\"\n",
       "output: \"1072\"\n",
       "name: \"Shape_558\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1072\"\n",
       "output: \"1073\"\n",
       "name: \"ConstantOfShape_559\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1074\"\n",
       "name: \"Constant_560\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1073\"\n",
       "input: \"1074\"\n",
       "output: \"1075\"\n",
       "name: \"Mul_561\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1063\"\n",
       "input: \"1075\"\n",
       "output: \"1076\"\n",
       "name: \"Equal_562\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1076\"\n",
       "input: \"1073\"\n",
       "input: \"1063\"\n",
       "output: \"1077\"\n",
       "name: \"Where_563\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1052\"\n",
       "input: \"1077\"\n",
       "output: \"1078\"\n",
       "name: \"Expand_564\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1078\"\n",
       "output: \"1079\"\n",
       "name: \"Unsqueeze_565\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1063\"\n",
       "output: \"1080\"\n",
       "name: \"Shape_566\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1080\"\n",
       "output: \"1081\"\n",
       "name: \"ConstantOfShape_567\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1082\"\n",
       "name: \"Constant_568\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1081\"\n",
       "input: \"1082\"\n",
       "output: \"1083\"\n",
       "name: \"Mul_569\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1063\"\n",
       "input: \"1083\"\n",
       "output: \"1084\"\n",
       "name: \"Equal_570\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1084\"\n",
       "input: \"1081\"\n",
       "input: \"1063\"\n",
       "output: \"1085\"\n",
       "name: \"Where_571\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1054\"\n",
       "input: \"1085\"\n",
       "output: \"1086\"\n",
       "name: \"Expand_572\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1086\"\n",
       "output: \"1087\"\n",
       "name: \"Unsqueeze_573\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1063\"\n",
       "output: \"1088\"\n",
       "name: \"Shape_574\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1088\"\n",
       "output: \"1089\"\n",
       "name: \"ConstantOfShape_575\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1090\"\n",
       "name: \"Constant_576\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1089\"\n",
       "input: \"1090\"\n",
       "output: \"1091\"\n",
       "name: \"Mul_577\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1063\"\n",
       "input: \"1091\"\n",
       "output: \"1092\"\n",
       "name: \"Equal_578\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1092\"\n",
       "input: \"1089\"\n",
       "input: \"1063\"\n",
       "output: \"1093\"\n",
       "name: \"Where_579\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1056\"\n",
       "input: \"1093\"\n",
       "output: \"1094\"\n",
       "name: \"Expand_580\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1094\"\n",
       "output: \"1095\"\n",
       "name: \"Unsqueeze_581\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1063\"\n",
       "output: \"1096\"\n",
       "name: \"Shape_582\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1096\"\n",
       "output: \"1097\"\n",
       "name: \"ConstantOfShape_583\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1098\"\n",
       "name: \"Constant_584\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1097\"\n",
       "input: \"1098\"\n",
       "output: \"1099\"\n",
       "name: \"Mul_585\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1063\"\n",
       "input: \"1099\"\n",
       "output: \"1100\"\n",
       "name: \"Equal_586\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1100\"\n",
       "input: \"1097\"\n",
       "input: \"1063\"\n",
       "output: \"1101\"\n",
       "name: \"Where_587\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1058\"\n",
       "input: \"1101\"\n",
       "output: \"1102\"\n",
       "name: \"Expand_588\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1102\"\n",
       "output: \"1103\"\n",
       "name: \"Unsqueeze_589\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1071\"\n",
       "input: \"1079\"\n",
       "input: \"1087\"\n",
       "input: \"1095\"\n",
       "input: \"1103\"\n",
       "output: \"1104\"\n",
       "name: \"Concat_590\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: -1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"990\"\n",
       "output: \"1105\"\n",
       "name: \"Shape_591\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1106\"\n",
       "name: \"Constant_592\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1107\"\n",
       "name: \"Constant_593\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\005\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1108\"\n",
       "name: \"Constant_594\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\177\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1105\"\n",
       "input: \"1107\"\n",
       "input: \"1108\"\n",
       "input: \"1106\"\n",
       "output: \"1109\"\n",
       "name: \"Slice_595\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"1063\"\n",
       "input: \"1109\"\n",
       "output: \"1110\"\n",
       "name: \"Concat_596\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1008\"\n",
       "input: \"1110\"\n",
       "output: \"1111\"\n",
       "name: \"Reshape_597\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"990\"\n",
       "input: \"1104\"\n",
       "input: \"1111\"\n",
       "output: \"1112\"\n",
       "name: \"ScatterND_598\"\n",
       "op_type: \"ScatterND\"\n",
       ", input: \"732\"\n",
       "output: \"1115\"\n",
       "name: \"Unsqueeze_599\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1115\"\n",
       "input: \"1543\"\n",
       "input: \"1544\"\n",
       "output: \"1118\"\n",
       "name: \"Concat_600\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1112\"\n",
       "input: \"1118\"\n",
       "output: \"1119\"\n",
       "name: \"Reshape_601\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1128\"\n",
       "name: \"Constant_602\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 0\n",
       "    data_type: 1\n",
       "    raw_data: \"\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"718\"\n",
       "input: \"1128\"\n",
       "input: \"1549\"\n",
       "output: \"1129\"\n",
       "name: \"Resize_603\"\n",
       "op_type: \"Resize\"\n",
       "attribute {\n",
       "  name: \"coordinate_transformation_mode\"\n",
       "  s: \"asymmetric\"\n",
       "  type: STRING\n",
       "}\n",
       "attribute {\n",
       "  name: \"cubic_coeff_a\"\n",
       "  f: -0.75\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"mode\"\n",
       "  s: \"nearest\"\n",
       "  type: STRING\n",
       "}\n",
       "attribute {\n",
       "  name: \"nearest_mode\"\n",
       "  s: \"floor\"\n",
       "  type: STRING\n",
       "}\n",
       ", input: \"1129\"\n",
       "input: \"665\"\n",
       "output: \"1130\"\n",
       "name: \"Concat_604\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1130\"\n",
       "input: \"module_list.119.Conv2d.weight\"\n",
       "output: \"1131\"\n",
       "name: \"Conv_605\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1131\"\n",
       "input: \"module_list.119.BatchNorm2d.weight\"\n",
       "input: \"module_list.119.BatchNorm2d.bias\"\n",
       "input: \"module_list.119.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.119.BatchNorm2d.running_var\"\n",
       "output: \"1132\"\n",
       "name: \"BatchNormalization_606\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1132\"\n",
       "output: \"1133\"\n",
       "name: \"LeakyRelu_607\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1133\"\n",
       "input: \"module_list.120.Conv2d.weight\"\n",
       "output: \"1134\"\n",
       "name: \"Conv_608\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 5\n",
       "  ints: 5\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1134\"\n",
       "input: \"module_list.120.BatchNorm2d.weight\"\n",
       "input: \"module_list.120.BatchNorm2d.bias\"\n",
       "input: \"module_list.120.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.120.BatchNorm2d.running_var\"\n",
       "output: \"1135\"\n",
       "name: \"BatchNormalization_609\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1135\"\n",
       "output: \"1136\"\n",
       "name: \"LeakyRelu_610\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1136\"\n",
       "input: \"module_list.121.Conv2d.weight\"\n",
       "output: \"1137\"\n",
       "name: \"Conv_611\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1137\"\n",
       "input: \"module_list.121.BatchNorm2d.weight\"\n",
       "input: \"module_list.121.BatchNorm2d.bias\"\n",
       "input: \"module_list.121.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.121.BatchNorm2d.running_var\"\n",
       "output: \"1138\"\n",
       "name: \"BatchNormalization_612\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1138\"\n",
       "input: \"module_list.122.Conv2d.weight\"\n",
       "output: \"1139\"\n",
       "name: \"Conv_613\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 96\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 5\n",
       "  ints: 5\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1139\"\n",
       "input: \"module_list.122.BatchNorm2d.weight\"\n",
       "input: \"module_list.122.BatchNorm2d.bias\"\n",
       "input: \"module_list.122.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.122.BatchNorm2d.running_var\"\n",
       "output: \"1140\"\n",
       "name: \"BatchNormalization_614\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1140\"\n",
       "output: \"1141\"\n",
       "name: \"LeakyRelu_615\"\n",
       "op_type: \"LeakyRelu\"\n",
       "attribute {\n",
       "  name: \"alpha\"\n",
       "  f: 0.10000000149011612\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1141\"\n",
       "input: \"module_list.123.Conv2d.weight\"\n",
       "output: \"1142\"\n",
       "name: \"Conv_616\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1142\"\n",
       "input: \"module_list.123.BatchNorm2d.weight\"\n",
       "input: \"module_list.123.BatchNorm2d.bias\"\n",
       "input: \"module_list.123.BatchNorm2d.running_mean\"\n",
       "input: \"module_list.123.BatchNorm2d.running_var\"\n",
       "output: \"1143\"\n",
       "name: \"BatchNormalization_617\"\n",
       "op_type: \"BatchNormalization\"\n",
       "attribute {\n",
       "  name: \"epsilon\"\n",
       "  f: 9.999999747378752e-05\n",
       "  type: FLOAT\n",
       "}\n",
       "attribute {\n",
       "  name: \"momentum\"\n",
       "  f: 0.9700000286102295\n",
       "  type: FLOAT\n",
       "}\n",
       ", input: \"1143\"\n",
       "input: \"module_list.124.Conv2d.weight\"\n",
       "input: \"module_list.124.Conv2d.bias\"\n",
       "output: \"1144\"\n",
       "name: \"Conv_618\"\n",
       "op_type: \"Conv\"\n",
       "attribute {\n",
       "  name: \"dilations\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"group\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "attribute {\n",
       "  name: \"kernel_shape\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"pads\"\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       "attribute {\n",
       "  name: \"strides\"\n",
       "  ints: 1\n",
       "  ints: 1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1144\"\n",
       "output: \"1145\"\n",
       "name: \"Shape_619\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1146\"\n",
       "name: \"Constant_620\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1145\"\n",
       "input: \"1146\"\n",
       "output: \"1147\"\n",
       "name: \"Gather_621\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1147\"\n",
       "output: \"1152\"\n",
       "name: \"Unsqueeze_622\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1152\"\n",
       "input: \"1550\"\n",
       "input: \"1551\"\n",
       "input: \"1552\"\n",
       "input: \"1553\"\n",
       "output: \"1157\"\n",
       "name: \"Concat_623\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1144\"\n",
       "input: \"1157\"\n",
       "output: \"1158\"\n",
       "name: \"Reshape_624\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1158\"\n",
       "output: \"yolo2\"\n",
       "name: \"Transpose_625\"\n",
       "op_type: \"Transpose\"\n",
       "attribute {\n",
       "  name: \"perm\"\n",
       "  ints: 0\n",
       "  ints: 1\n",
       "  ints: 3\n",
       "  ints: 4\n",
       "  ints: 2\n",
       "  type: INTS\n",
       "}\n",
       ", output: \"1160\"\n",
       "name: \"Constant_626\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1161\"\n",
       "name: \"Constant_627\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1162\"\n",
       "name: \"Constant_628\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1163\"\n",
       "name: \"Constant_629\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"yolo2\"\n",
       "input: \"1161\"\n",
       "input: \"1162\"\n",
       "input: \"1160\"\n",
       "input: \"1163\"\n",
       "output: \"1164\"\n",
       "name: \"Slice_630\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"1164\"\n",
       "output: \"1165\"\n",
       "name: \"Sigmoid_631\"\n",
       "op_type: \"Sigmoid\"\n",
       ", output: \"1166\"\n",
       "name: \"Constant_632\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    dims: 1\n",
       "    dims: 20\n",
       "    dims: 32\n",
       "    dims: 2\n",
       "    data_type: 1\n",
       "    raw_data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`A\\000\\000\\000\\000\\000\\000pA\\000\\000\\000\\000\\000\\000\\200A\\000\\000\\000\\000\\000\\000\\210A\\000\\000\\000\\000\\000\\000\\220A\\000\\000\\000\\000\\000\\000\\230A\\000\\000\\000\\000\\000\\000\\240A\\000\\000\\000\\000\\000\\000\\250A\\000\\000\\000\\000\\000\\000\\260A\\000\\000\\000\\000\\000\\000\\270A\\000\\000\\000\\000\\000\\000\\300A\\000\\000\\000\\000\\000\\000\\310A\\000\\000\\000\\000\\000\\000\\320A\\000\\000\\000\\000\\000\\000\\330A\\000\\000\\000\\000\\000\\000\\340A\\000\\000\\000\\000\\000\\000\\350A\\000\\000\\000\\000\\000\\000\\360A\\000\\000\\000\\000\\000\\000\\370A\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\200?\\000\\000\\200?\\000\\000\\200?\\000\\000\\000@\\000\\000\\200?\\000\\000@@\\000\\000\\200?\\000\\000\\200@\\000\\000\\200?\\000\\000\\240@\\000\\000\\200?\\000\\000\\300@\\000\\000\\200?\\000\\000\\340@\\000\\000\\200?\\000\\000\\000A\\000\\000\\200?\\000\\000\\020A\\000\\000\\200?\\000\\000 A\\000\\000\\200?\\000\\0000A\\000\\000\\200?\\000\\000@A\\000\\000\\200?\\000\\000PA\\000\\000\\200?\\000\\000`A\\000\\000\\200?\\000\\000pA\\000\\000\\200?\\000\\000\\200A\\000\\000\\200?\\000\\000\\210A\\000\\000\\200?\\000\\000\\220A\\000\\000\\200?\\000\\000\\230A\\000\\000\\200?\\000\\000\\240A\\000\\000\\200?\\000\\000\\250A\\000\\000\\200?\\000\\000\\260A\\000\\000\\200?\\000\\000\\270A\\000\\000\\200?\\000\\000\\300A\\000\\000\\200?\\000\\000\\310A\\000\\000\\200?\\000\\000\\320A\\000\\000\\200?\\000\\000\\330A\\000\\000\\200?\\000\\000\\340A\\000\\000\\200?\\000\\000\\350A\\000\\000\\200?\\000\\000\\360A\\000\\000\\200?\\000\\000\\370A\\000\\000\\200?\\000\\000\\000\\000\\000\\000\\000@\\000\\000\\200?\\000\\000\\000@\\000\\000\\000@\\000\\000\\000@\\000\\000@@\\000\\000\\000@\\000\\000\\200@\\000\\000\\000@\\000\\000\\240@\\000\\000\\000@\\000\\000\\300@\\000\\000\\000@\\000\\000\\340@\\000\\000\\000@\\000\\000\\000A\\000\\000\\000@\\000\\000\\020A\\000\\000\\000@\\000\\000 A\\000\\000\\000@\\000\\0000A\\000\\000\\000@\\000\\000@A\\000\\000\\000@\\000\\000PA\\000\\000\\000@\\000\\000`A\\000\\000\\000@\\000\\000pA\\000\\000\\000@\\000\\000\\200A\\000\\000\\000@\\000\\000\\210A\\000\\000\\000@\\000\\000\\220A\\000\\000\\000@\\000\\000\\230A\\000\\000\\000@\\000\\000\\240A\\000\\000\\000@\\000\\000\\250A\\000\\000\\000@\\000\\000\\260A\\000\\000\\000@\\000\\000\\270A\\000\\000\\000@\\000\\000\\300A\\000\\000\\000@\\000\\000\\310A\\000\\000\\000@\\000\\000\\320A\\000\\000\\000@\\000\\000\\330A\\000\\000\\000@\\000\\000\\340A\\000\\000\\000@\\000\\000\\350A\\000\\000\\000@\\000\\000\\360A\\000\\000\\000@\\000\\000\\370A\\000\\000\\000@\\000\\000\\000\\000\\000\\000@@\\000\\000\\200?\\000\\000@@\\000\\000\\000@\\000\\000@@\\000\\000@@\\000\\000@@\\000\\000\\200@\\000\\000@@\\000\\000\\240@\\000\\000@@\\000\\000\\300@\\000\\000@@\\000\\000\\340@\\000\\000@@\\000\\000\\000A\\000\\000@@\\000\\000\\020A\\000\\000@@\\000\\000 A\\000\\000@@\\000\\0000A\\000\\000@@\\000\\000@A\\000\\000@@\\000\\000PA\\000\\000@@\\000\\000`A\\000\\000@@\\000\\000pA\\000\\000@@\\000\\000\\200A\\000\\000@@\\000\\000\\210A\\000\\000@@\\000\\000\\220A\\000\\000@@\\000\\000\\230A\\000\\000@@\\000\\000\\240A\\000\\000@@\\000\\000\\250A\\000\\000@@\\000\\000\\260A\\000\\000@@\\000\\000\\270A\\000\\000@@\\000\\000\\300A\\000\\000@@\\000\\000\\310A\\000\\000@@\\000\\000\\320A\\000\\000@@\\000\\000\\330A\\000\\000@@\\000\\000\\340A\\000\\000@@\\000\\000\\350A\\000\\000@@\\000\\000\\360A\\000\\000@@\\000\\000\\370A\\000\\000@@\\000\\000\\000\\000\\000\\000\\200@\\000\\000\\200?\\000\\000\\200@\\000\\000\\000@\\000\\000\\200@\\000\\000@@\\000\\000\\200@\\000\\000\\200@\\000\\000\\200@\\000\\000\\240@\\000\\000\\200@\\000\\000\\300@\\000\\000\\200@\\000\\000\\340@\\000\\000\\200@\\000\\000\\000A\\000\\000\\200@\\000\\000\\020A\\000\\000\\200@\\000\\000 A\\000\\000\\200@\\000\\0000A\\000\\000\\200@\\000\\000@A\\000\\000\\200@\\000\\000PA\\000\\000\\200@\\000\\000`A\\000\\000\\200@\\000\\000pA\\000\\000\\200@\\000\\000\\200A\\000\\000\\200@\\000\\000\\210A\\000\\000\\200@\\000\\000\\220A\\000\\000\\200@\\000\\000\\230A\\000\\000\\200@\\000\\000\\240A\\000\\000\\200@\\000\\000\\250A\\000\\000\\200@\\000\\000\\260A\\000\\000\\200@\\000\\000\\270A\\000\\000\\200@\\000\\000\\300A\\000\\000\\200@\\000\\000\\310A\\000\\000\\200@\\000\\000\\320A\\000\\000\\200@\\000\\000\\330A\\000\\000\\200@\\000\\000\\340A\\000\\000\\200@\\000\\000\\350A\\000\\000\\200@\\000\\000\\360A\\000\\000\\200@\\000\\000\\370A\\000\\000\\200@\\000\\000\\000\\000\\000\\000\\240@\\000\\000\\200?\\000\\000\\240@\\000\\000\\000@\\000\\000\\240@\\000\\000@@\\000\\000\\240@\\000\\000\\200@\\000\\000\\240@\\000\\000\\240@\\000\\000\\240@\\000\\000\\300@\\000\\000\\240@\\000\\000\\340@\\000\\000\\240@\\000\\000\\000A\\000\\000\\240@\\000\\000\\020A\\000\\000\\240@\\000\\000 A\\000\\000\\240@\\000\\0000A\\000\\000\\240@\\000\\000@A\\000\\000\\240@\\000\\000PA\\000\\000\\240@\\000\\000`A\\000\\000\\240@\\000\\000pA\\000\\000\\240@\\000\\000\\200A\\000\\000\\240@\\000\\000\\210A\\000\\000\\240@\\000\\000\\220A\\000\\000\\240@\\000\\000\\230A\\000\\000\\240@\\000\\000\\240A\\000\\000\\240@\\000\\000\\250A\\000\\000\\240@\\000\\000\\260A\\000\\000\\240@\\000\\000\\270A\\000\\000\\240@\\000\\000\\300A\\000\\000\\240@\\000\\000\\310A\\000\\000\\240@\\000\\000\\320A\\000\\000\\240@\\000\\000\\330A\\000\\000\\240@\\000\\000\\340A\\000\\000\\240@\\000\\000\\350A\\000\\000\\240@\\000\\000\\360A\\000\\000\\240@\\000\\000\\370A\\000\\000\\240@\\000\\000\\000\\000\\000\\000\\300@\\000\\000\\200?\\000\\000\\300@\\000\\000\\000@\\000\\000\\300@\\000\\000@@\\000\\000\\300@\\000\\000\\200@\\000\\000\\300@\\000\\000\\240@\\000\\000\\300@\\000\\000\\300@\\000\\000\\300@\\000\\000\\340@\\000\\000\\300@\\000\\000\\000A\\000\\000\\300@\\000\\000\\020A\\000\\000\\300@\\000\\000 A\\000\\000\\300@\\000\\0000A\\000\\000\\300@\\000\\000@A\\000\\000\\300@\\000\\000PA\\000\\000\\300@\\000\\000`A\\000\\000\\300@\\000\\000pA\\000\\000\\300@\\000\\000\\200A\\000\\000\\300@\\000\\000\\210A\\000\\000\\300@\\000\\000\\220A\\000\\000\\300@\\000\\000\\230A\\000\\000\\300@\\000\\000\\240A\\000\\000\\300@\\000\\000\\250A\\000\\000\\300@\\000\\000\\260A\\000\\000\\300@\\000\\000\\270A\\000\\000\\300@\\000\\000\\300A\\000\\000\\300@\\000\\000\\310A\\000\\000\\300@\\000\\000\\320A\\000\\000\\300@\\000\\000\\330A\\000\\000\\300@\\000\\000\\340A\\000\\000\\300@\\000\\000\\350A\\000\\000\\300@\\000\\000\\360A\\000\\000\\300@\\000\\000\\370A\\000\\000\\300@\\000\\000\\000\\000\\000\\000\\340@\\000\\000\\200?\\000\\000\\340@\\000\\000\\000@\\000\\000\\340@\\000\\000@@\\000\\000\\340@\\000\\000\\200@\\000\\000\\340@\\000\\000\\240@\\000\\000\\340@\\000\\000\\300@\\000\\000\\340@\\000\\000\\340@\\000\\000\\340@\\000\\000\\000A\\000\\000\\340@\\000\\000\\020A\\000\\000\\340@\\000\\000 A\\000\\000\\340@\\000\\0000A\\000\\000\\340@\\000\\000@A\\000\\000\\340@\\000\\000PA\\000\\000\\340@\\000\\000`A\\000\\000\\340@\\000\\000pA\\000\\000\\340@\\000\\000\\200A\\000\\000\\340@\\000\\000\\210A\\000\\000\\340@\\000\\000\\220A\\000\\000\\340@\\000\\000\\230A\\000\\000\\340@\\000\\000\\240A\\000\\000\\340@\\000\\000\\250A\\000\\000\\340@\\000\\000\\260A\\000\\000\\340@\\000\\000\\270A\\000\\000\\340@\\000\\000\\300A\\000\\000\\340@\\000\\000\\310A\\000\\000\\340@\\000\\000\\320A\\000\\000\\340@\\000\\000\\330A\\000\\000\\340@\\000\\000\\340A\\000\\000\\340@\\000\\000\\350A\\000\\000\\340@\\000\\000\\360A\\000\\000\\340@\\000\\000\\370A\\000\\000\\340@\\000\\000\\000\\000\\000\\000\\000A\\000\\000\\200?\\000\\000\\000A\\000\\000\\000@\\000\\000\\000A\\000\\000@@\\000\\000\\000A\\000\\000\\200@\\000\\000\\000A\\000\\000\\240@\\000\\000\\000A\\000\\000\\300@\\000\\000\\000A\\000\\000\\340@\\000\\000\\000A\\000\\000\\000A\\000\\000\\000A\\000\\000\\020A\\000\\000\\000A\\000\\000 A\\000\\000\\000A\\000\\0000A\\000\\000\\000A\\000\\000@A\\000\\000\\000A\\000\\000PA\\000\\000\\000A\\000\\000`A\\000\\000\\000A\\000\\000pA\\000\\000\\000A\\000\\000\\200A\\000\\000\\000A\\000\\000\\210A\\000\\000\\000A\\000\\000\\220A\\000\\000\\000A\\000\\000\\230A\\000\\000\\000A\\000\\000\\240A\\000\\000\\000A\\000\\000\\250A\\000\\000\\000A\\000\\000\\260A\\000\\000\\000A\\000\\000\\270A\\000\\000\\000A\\000\\000\\300A\\000\\000\\000A\\000\\000\\310A\\000\\000\\000A\\000\\000\\320A\\000\\000\\000A\\000\\000\\330A\\000\\000\\000A\\000\\000\\340A\\000\\000\\000A\\000\\000\\350A\\000\\000\\000A\\000\\000\\360A\\000\\000\\000A\\000\\000\\370A\\000\\000\\000A\\000\\000\\000\\000\\000\\000\\020A\\000\\000\\200?\\000\\000\\020A\\000\\000\\000@\\000\\000\\020A\\000\\000@@\\000\\000\\020A\\000\\000\\200@\\000\\000\\020A\\000\\000\\240@\\000\\000\\020A\\000\\000\\300@\\000\\000\\020A\\000\\000\\340@\\000\\000\\020A\\000\\000\\000A\\000\\000\\020A\\000\\000\\020A\\000\\000\\020A\\000\\000 A\\000\\000\\020A\\000\\0000A\\000\\000\\020A\\000\\000@A\\000\\000\\020A\\000\\000PA\\000\\000\\020A\\000\\000`A\\000\\000\\020A\\000\\000pA\\000\\000\\020A\\000\\000\\200A\\000\\000\\020A\\000\\000\\210A\\000\\000\\020A\\000\\000\\220A\\000\\000\\020A\\000\\000\\230A\\000\\000\\020A\\000\\000\\240A\\000\\000\\020A\\000\\000\\250A\\000\\000\\020A\\000\\000\\260A\\000\\000\\020A\\000\\000\\270A\\000\\000\\020A\\000\\000\\300A\\000\\000\\020A\\000\\000\\310A\\000\\000\\020A\\000\\000\\320A\\000\\000\\020A\\000\\000\\330A\\000\\000\\020A\\000\\000\\340A\\000\\000\\020A\\000\\000\\350A\\000\\000\\020A\\000\\000\\360A\\000\\000\\020A\\000\\000\\370A\\000\\000\\020A\\000\\000\\000\\000\\000\\000 A\\000\\000\\200?\\000\\000 A\\000\\000\\000@\\000\\000 A\\000\\000@@\\000\\000 A\\000\\000\\200@\\000\\000 A\\000\\000\\240@\\000\\000 A\\000\\000\\300@\\000\\000 A\\000\\000\\340@\\000\\000 A\\000\\000\\000A\\000\\000 A\\000\\000\\020A\\000\\000 A\\000\\000 A\\000\\000 A\\000\\0000A\\000\\000 A\\000\\000@A\\000\\000 A\\000\\000PA\\000\\000 A\\000\\000`A\\000\\000 A\\000\\000pA\\000\\000 A\\000\\000\\200A\\000\\000 A\\000\\000\\210A\\000\\000 A\\000\\000\\220A\\000\\000 A\\000\\000\\230A\\000\\000 A\\000\\000\\240A\\000\\000 A\\000\\000\\250A\\000\\000 A\\000\\000\\260A\\000\\000 A\\000\\000\\270A\\000\\000 A\\000\\000\\300A\\000\\000 A\\000\\000\\310A\\000\\000 A\\000\\000\\320A\\000\\000 A\\000\\000\\330A\\000\\000 A\\000\\000\\340A\\000\\000 A\\000\\000\\350A\\000\\000 A\\000\\000\\360A\\000\\000 A\\000\\000\\370A\\000\\000 A\\000\\000\\000\\000\\000\\0000A\\000\\000\\200?\\000\\0000A\\000\\000\\000@\\000\\0000A\\000\\000@@\\000\\0000A\\000\\000\\200@\\000\\0000A\\000\\000\\240@\\000\\0000A\\000\\000\\300@\\000\\0000A\\000\\000\\340@\\000\\0000A\\000\\000\\000A\\000\\0000A\\000\\000\\020A\\000\\0000A\\000\\000 A\\000\\0000A\\000\\0000A\\000\\0000A\\000\\000@A\\000\\0000A\\000\\000PA\\000\\0000A\\000\\000`A\\000\\0000A\\000\\000pA\\000\\0000A\\000\\000\\200A\\000\\0000A\\000\\000\\210A\\000\\0000A\\000\\000\\220A\\000\\0000A\\000\\000\\230A\\000\\0000A\\000\\000\\240A\\000\\0000A\\000\\000\\250A\\000\\0000A\\000\\000\\260A\\000\\0000A\\000\\000\\270A\\000\\0000A\\000\\000\\300A\\000\\0000A\\000\\000\\310A\\000\\0000A\\000\\000\\320A\\000\\0000A\\000\\000\\330A\\000\\0000A\\000\\000\\340A\\000\\0000A\\000\\000\\350A\\000\\0000A\\000\\000\\360A\\000\\0000A\\000\\000\\370A\\000\\0000A\\000\\000\\000\\000\\000\\000@A\\000\\000\\200?\\000\\000@A\\000\\000\\000@\\000\\000@A\\000\\000@@\\000\\000@A\\000\\000\\200@\\000\\000@A\\000\\000\\240@\\000\\000@A\\000\\000\\300@\\000\\000@A\\000\\000\\340@\\000\\000@A\\000\\000\\000A\\000\\000@A\\000\\000\\020A\\000\\000@A\\000\\000 A\\000\\000@A\\000\\0000A\\000\\000@A\\000\\000@A\\000\\000@A\\000\\000PA\\000\\000@A\\000\\000`A\\000\\000@A\\000\\000pA\\000\\000@A\\000\\000\\200A\\000\\000@A\\000\\000\\210A\\000\\000@A\\000\\000\\220A\\000\\000@A\\000\\000\\230A\\000\\000@A\\000\\000\\240A\\000\\000@A\\000\\000\\250A\\000\\000@A\\000\\000\\260A\\000\\000@A\\000\\000\\270A\\000\\000@A\\000\\000\\300A\\000\\000@A\\000\\000\\310A\\000\\000@A\\000\\000\\320A\\000\\000@A\\000\\000\\330A\\000\\000@A\\000\\000\\340A\\000\\000@A\\000\\000\\350A\\000\\000@A\\000\\000\\360A\\000\\000@A\\000\\000\\370A\\000\\000@A\\000\\000\\000\\000\\000\\000PA\\000\\000\\200?\\000\\000PA\\000\\000\\000@\\000\\000PA\\000\\000@@\\000\\000PA\\000\\000\\200@\\000\\000PA\\000\\000\\240@\\000\\000PA\\000\\000\\300@\\000\\000PA\\000\\000\\340@\\000\\000PA\\000\\000\\000A\\000\\000PA\\000\\000\\020A\\000\\000PA\\000\\000 A\\000\\000PA\\000\\0000A\\000\\000PA\\000\\000@A\\000\\000PA\\000\\000PA\\000\\000PA\\000\\000`A\\000\\000PA\\000\\000pA\\000\\000PA\\000\\000\\200A\\000\\000PA\\000\\000\\210A\\000\\000PA\\000\\000\\220A\\000\\000PA\\000\\000\\230A\\000\\000PA\\000\\000\\240A\\000\\000PA\\000\\000\\250A\\000\\000PA\\000\\000\\260A\\000\\000PA\\000\\000\\270A\\000\\000PA\\000\\000\\300A\\000\\000PA\\000\\000\\310A\\000\\000PA\\000\\000\\320A\\000\\000PA\\000\\000\\330A\\000\\000PA\\000\\000\\340A\\000\\000PA\\000\\000\\350A\\000\\000PA\\000\\000\\360A\\000\\000PA\\000\\000\\370A\\000\\000PA\\000\\000\\000\\000\\000\\000`A\\000\\000\\200?\\000\\000`A\\000\\000\\000@\\000\\000`A\\000\\000@@\\000\\000`A\\000\\000\\200@\\000\\000`A\\000\\000\\240@\\000\\000`A\\000\\000\\300@\\000\\000`A\\000\\000\\340@\\000\\000`A\\000\\000\\000A\\000\\000`A\\000\\000\\020A\\000\\000`A\\000\\000 A\\000\\000`A\\000\\0000A\\000\\000`A\\000\\000@A\\000\\000`A\\000\\000PA\\000\\000`A\\000\\000`A\\000\\000`A\\000\\000pA\\000\\000`A\\000\\000\\200A\\000\\000`A\\000\\000\\210A\\000\\000`A\\000\\000\\220A\\000\\000`A\\000\\000\\230A\\000\\000`A\\000\\000\\240A\\000\\000`A\\000\\000\\250A\\000\\000`A\\000\\000\\260A\\000\\000`A\\000\\000\\270A\\000\\000`A\\000\\000\\300A\\000\\000`A\\000\\000\\310A\\000\\000`A\\000\\000\\320A\\000\\000`A\\000\\000\\330A\\000\\000`A\\000\\000\\340A\\000\\000`A\\000\\000\\350A\\000\\000`A\\000\\000\\360A\\000\\000`A\\000\\000\\370A\\000\\000`A\\000\\000\\000\\000\\000\\000pA\\000\\000\\200?\\000\\000pA\\000\\000\\000@\\000\\000pA\\000\\000@@\\000\\000pA\\000\\000\\200@\\000\\000pA\\000\\000\\240@\\000\\000pA\\000\\000\\300@\\000\\000pA\\000\\000\\340@\\000\\000pA\\000\\000\\000A\\000\\000pA\\000\\000\\020A\\000\\000pA\\000\\000 A\\000\\000pA\\000\\0000A\\000\\000pA\\000\\000@A\\000\\000pA\\000\\000PA\\000\\000pA\\000\\000`A\\000\\000pA\\000\\000pA\\000\\000pA\\000\\000\\200A\\000\\000pA\\000\\000\\210A\\000\\000pA\\000\\000\\220A\\000\\000pA\\000\\000\\230A\\000\\000pA\\000\\000\\240A\\000\\000pA\\000\\000\\250A\\000\\000pA\\000\\000\\260A\\000\\000pA\\000\\000\\270A\\000\\000pA\\000\\000\\300A\\000\\000pA\\000\\000\\310A\\000\\000pA\\000\\000\\320A\\000\\000pA\\000\\000\\330A\\000\\000pA\\000\\000\\340A\\000\\000pA\\000\\000\\350A\\000\\000pA\\000\\000\\360A\\000\\000pA\\000\\000\\370A\\000\\000pA\\000\\000\\000\\000\\000\\000\\200A\\000\\000\\200?\\000\\000\\200A\\000\\000\\000@\\000\\000\\200A\\000\\000@@\\000\\000\\200A\\000\\000\\200@\\000\\000\\200A\\000\\000\\240@\\000\\000\\200A\\000\\000\\300@\\000\\000\\200A\\000\\000\\340@\\000\\000\\200A\\000\\000\\000A\\000\\000\\200A\\000\\000\\020A\\000\\000\\200A\\000\\000 A\\000\\000\\200A\\000\\0000A\\000\\000\\200A\\000\\000@A\\000\\000\\200A\\000\\000PA\\000\\000\\200A\\000\\000`A\\000\\000\\200A\\000\\000pA\\000\\000\\200A\\000\\000\\200A\\000\\000\\200A\\000\\000\\210A\\000\\000\\200A\\000\\000\\220A\\000\\000\\200A\\000\\000\\230A\\000\\000\\200A\\000\\000\\240A\\000\\000\\200A\\000\\000\\250A\\000\\000\\200A\\000\\000\\260A\\000\\000\\200A\\000\\000\\270A\\000\\000\\200A\\000\\000\\300A\\000\\000\\200A\\000\\000\\310A\\000\\000\\200A\\000\\000\\320A\\000\\000\\200A\\000\\000\\330A\\000\\000\\200A\\000\\000\\340A\\000\\000\\200A\\000\\000\\350A\\000\\000\\200A\\000\\000\\360A\\000\\000\\200A\\000\\000\\370A\\000\\000\\200A\\000\\000\\000\\000\\000\\000\\210A\\000\\000\\200?\\000\\000\\210A\\000\\000\\000@\\000\\000\\210A\\000\\000@@\\000\\000\\210A\\000\\000\\200@\\000\\000\\210A\\000\\000\\240@\\000\\000\\210A\\000\\000\\300@\\000\\000\\210A\\000\\000\\340@\\000\\000\\210A\\000\\000\\000A\\000\\000\\210A\\000\\000\\020A\\000\\000\\210A\\000\\000 A\\000\\000\\210A\\000\\0000A\\000\\000\\210A\\000\\000@A\\000\\000\\210A\\000\\000PA\\000\\000\\210A\\000\\000`A\\000\\000\\210A\\000\\000pA\\000\\000\\210A\\000\\000\\200A\\000\\000\\210A\\000\\000\\210A\\000\\000\\210A\\000\\000\\220A\\000\\000\\210A\\000\\000\\230A\\000\\000\\210A\\000\\000\\240A\\000\\000\\210A\\000\\000\\250A\\000\\000\\210A\\000\\000\\260A\\000\\000\\210A\\000\\000\\270A\\000\\000\\210A\\000\\000\\300A\\000\\000\\210A\\000\\000\\310A\\000\\000\\210A\\000\\000\\320A\\000\\000\\210A\\000\\000\\330A\\000\\000\\210A\\000\\000\\340A\\000\\000\\210A\\000\\000\\350A\\000\\000\\210A\\000\\000\\360A\\000\\000\\210A\\000\\000\\370A\\000\\000\\210A\\000\\000\\000\\000\\000\\000\\220A\\000\\000\\200?\\000\\000\\220A\\000\\000\\000@\\000\\000\\220A\\000\\000@@\\000\\000\\220A\\000\\000\\200@\\000\\000\\220A\\000\\000\\240@\\000\\000\\220A\\000\\000\\300@\\000\\000\\220A\\000\\000\\340@\\000\\000\\220A\\000\\000\\000A\\000\\000\\220A\\000\\000\\020A\\000\\000\\220A\\000\\000 A\\000\\000\\220A\\000\\0000A\\000\\000\\220A\\000\\000@A\\000\\000\\220A\\000\\000PA\\000\\000\\220A\\000\\000`A\\000\\000\\220A\\000\\000pA\\000\\000\\220A\\000\\000\\200A\\000\\000\\220A\\000\\000\\210A\\000\\000\\220A\\000\\000\\220A\\000\\000\\220A\\000\\000\\230A\\000\\000\\220A\\000\\000\\240A\\000\\000\\220A\\000\\000\\250A\\000\\000\\220A\\000\\000\\260A\\000\\000\\220A\\000\\000\\270A\\000\\000\\220A\\000\\000\\300A\\000\\000\\220A\\000\\000\\310A\\000\\000\\220A\\000\\000\\320A\\000\\000\\220A\\000\\000\\330A\\000\\000\\220A\\000\\000\\340A\\000\\000\\220A\\000\\000\\350A\\000\\000\\220A\\000\\000\\360A\\000\\000\\220A\\000\\000\\370A\\000\\000\\220A\\000\\000\\000\\000\\000\\000\\230A\\000\\000\\200?\\000\\000\\230A\\000\\000\\000@\\000\\000\\230A\\000\\000@@\\000\\000\\230A\\000\\000\\200@\\000\\000\\230A\\000\\000\\240@\\000\\000\\230A\\000\\000\\300@\\000\\000\\230A\\000\\000\\340@\\000\\000\\230A\\000\\000\\000A\\000\\000\\230A\\000\\000\\020A\\000\\000\\230A\\000\\000 A\\000\\000\\230A\\000\\0000A\\000\\000\\230A\\000\\000@A\\000\\000\\230A\\000\\000PA\\000\\000\\230A\\000\\000`A\\000\\000\\230A\\000\\000pA\\000\\000\\230A\\000\\000\\200A\\000\\000\\230A\\000\\000\\210A\\000\\000\\230A\\000\\000\\220A\\000\\000\\230A\\000\\000\\230A\\000\\000\\230A\\000\\000\\240A\\000\\000\\230A\\000\\000\\250A\\000\\000\\230A\\000\\000\\260A\\000\\000\\230A\\000\\000\\270A\\000\\000\\230A\\000\\000\\300A\\000\\000\\230A\\000\\000\\310A\\000\\000\\230A\\000\\000\\320A\\000\\000\\230A\\000\\000\\330A\\000\\000\\230A\\000\\000\\340A\\000\\000\\230A\\000\\000\\350A\\000\\000\\230A\\000\\000\\360A\\000\\000\\230A\\000\\000\\370A\\000\\000\\230A\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1165\"\n",
       "input: \"1166\"\n",
       "output: \"1167\"\n",
       "name: \"Add_633\"\n",
       "op_type: \"Add\"\n",
       ", output: \"1168\"\n",
       "name: \"Constant_634\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1167\"\n",
       "input: \"1168\"\n",
       "output: \"1169\"\n",
       "name: \"Reshape_635\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1170\"\n",
       "name: \"Constant_636\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1554\"\n",
       "output: \"1172\"\n",
       "name: \"ConstantOfShape_637\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1173\"\n",
       "name: \"Constant_638\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1172\"\n",
       "input: \"1173\"\n",
       "output: \"1174\"\n",
       "name: \"Mul_639\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"1175\"\n",
       "name: \"Constant_640\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1175\"\n",
       "input: \"1174\"\n",
       "output: \"1176\"\n",
       "name: \"Equal_641\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1176\"\n",
       "input: \"1172\"\n",
       "input: \"1170\"\n",
       "output: \"1177\"\n",
       "name: \"Where_642\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1169\"\n",
       "input: \"1177\"\n",
       "output: \"1178\"\n",
       "name: \"Expand_643\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"yolo2\"\n",
       "output: \"1179\"\n",
       "name: \"Shape_644\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1180\"\n",
       "name: \"Constant_645\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1179\"\n",
       "input: \"1180\"\n",
       "output: \"1181\"\n",
       "name: \"Gather_646\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1181\"\n",
       "output: \"1182\"\n",
       "name: \"Cast_647\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1183\"\n",
       "name: \"Constant_648\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1184\"\n",
       "name: \"Constant_649\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1183\"\n",
       "input: \"1182\"\n",
       "input: \"1184\"\n",
       "output: \"1185\"\n",
       "name: \"Range_650\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo2\"\n",
       "output: \"1186\"\n",
       "name: \"Shape_651\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1187\"\n",
       "name: \"Constant_652\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1186\"\n",
       "input: \"1187\"\n",
       "output: \"1188\"\n",
       "name: \"Gather_653\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1188\"\n",
       "output: \"1189\"\n",
       "name: \"Cast_654\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1190\"\n",
       "name: \"Constant_655\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1191\"\n",
       "name: \"Constant_656\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1190\"\n",
       "input: \"1189\"\n",
       "input: \"1191\"\n",
       "output: \"1192\"\n",
       "name: \"Range_657\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo2\"\n",
       "output: \"1193\"\n",
       "name: \"Shape_658\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1194\"\n",
       "name: \"Constant_659\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1193\"\n",
       "input: \"1194\"\n",
       "output: \"1195\"\n",
       "name: \"Gather_660\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1195\"\n",
       "output: \"1196\"\n",
       "name: \"Cast_661\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1197\"\n",
       "name: \"Constant_662\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1198\"\n",
       "name: \"Constant_663\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1197\"\n",
       "input: \"1196\"\n",
       "input: \"1198\"\n",
       "output: \"1199\"\n",
       "name: \"Range_664\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo2\"\n",
       "output: \"1200\"\n",
       "name: \"Shape_665\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1201\"\n",
       "name: \"Constant_666\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1200\"\n",
       "input: \"1201\"\n",
       "output: \"1202\"\n",
       "name: \"Gather_667\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1202\"\n",
       "output: \"1203\"\n",
       "name: \"Cast_668\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1204\"\n",
       "name: \"Constant_669\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1205\"\n",
       "name: \"Constant_670\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1204\"\n",
       "input: \"1203\"\n",
       "input: \"1205\"\n",
       "output: \"1206\"\n",
       "name: \"Range_671\"\n",
       "op_type: \"Range\"\n",
       ", input: \"yolo2\"\n",
       "output: \"1207\"\n",
       "name: \"Shape_672\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1208\"\n",
       "name: \"Constant_673\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1207\"\n",
       "input: \"1208\"\n",
       "output: \"1209\"\n",
       "name: \"Gather_674\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1209\"\n",
       "output: \"1210\"\n",
       "name: \"Cast_675\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1211\"\n",
       "name: \"Constant_676\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1212\"\n",
       "name: \"Constant_677\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1211\"\n",
       "input: \"1210\"\n",
       "input: \"1212\"\n",
       "output: \"1213\"\n",
       "name: \"Range_678\"\n",
       "op_type: \"Range\"\n",
       ", output: \"1214\"\n",
       "name: \"Constant_679\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1215\"\n",
       "name: \"Constant_680\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1216\"\n",
       "name: \"Constant_681\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1217\"\n",
       "name: \"Constant_682\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1213\"\n",
       "input: \"1215\"\n",
       "input: \"1216\"\n",
       "input: \"1214\"\n",
       "input: \"1217\"\n",
       "output: \"1218\"\n",
       "name: \"Slice_683\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"1219\"\n",
       "name: \"Constant_684\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1185\"\n",
       "input: \"1219\"\n",
       "output: \"1220\"\n",
       "name: \"Reshape_685\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1221\"\n",
       "name: \"Constant_686\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1192\"\n",
       "input: \"1221\"\n",
       "output: \"1222\"\n",
       "name: \"Reshape_687\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1223\"\n",
       "name: \"Constant_688\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 3\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1199\"\n",
       "input: \"1223\"\n",
       "output: \"1224\"\n",
       "name: \"Reshape_689\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1225\"\n",
       "name: \"Constant_690\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 2\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1206\"\n",
       "input: \"1225\"\n",
       "output: \"1226\"\n",
       "name: \"Reshape_691\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1227\"\n",
       "name: \"Constant_692\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1218\"\n",
       "input: \"1227\"\n",
       "output: \"1228\"\n",
       "name: \"Reshape_693\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1220\"\n",
       "input: \"1222\"\n",
       "output: \"1229\"\n",
       "name: \"Add_694\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1229\"\n",
       "input: \"1224\"\n",
       "output: \"1230\"\n",
       "name: \"Add_695\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1230\"\n",
       "input: \"1226\"\n",
       "output: \"1231\"\n",
       "name: \"Add_696\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1231\"\n",
       "input: \"1228\"\n",
       "output: \"1232\"\n",
       "name: \"Add_697\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1232\"\n",
       "output: \"1233\"\n",
       "name: \"Shape_698\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1233\"\n",
       "output: \"1234\"\n",
       "name: \"Shape_699\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1234\"\n",
       "output: \"1235\"\n",
       "name: \"ConstantOfShape_700\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1236\"\n",
       "name: \"Constant_701\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1235\"\n",
       "input: \"1236\"\n",
       "output: \"1237\"\n",
       "name: \"Mul_702\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1233\"\n",
       "input: \"1237\"\n",
       "output: \"1238\"\n",
       "name: \"Equal_703\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1238\"\n",
       "input: \"1235\"\n",
       "input: \"1233\"\n",
       "output: \"1239\"\n",
       "name: \"Where_704\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1220\"\n",
       "input: \"1239\"\n",
       "output: \"1240\"\n",
       "name: \"Expand_705\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1240\"\n",
       "output: \"1241\"\n",
       "name: \"Unsqueeze_706\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1233\"\n",
       "output: \"1242\"\n",
       "name: \"Shape_707\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1242\"\n",
       "output: \"1243\"\n",
       "name: \"ConstantOfShape_708\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1244\"\n",
       "name: \"Constant_709\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1243\"\n",
       "input: \"1244\"\n",
       "output: \"1245\"\n",
       "name: \"Mul_710\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1233\"\n",
       "input: \"1245\"\n",
       "output: \"1246\"\n",
       "name: \"Equal_711\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1246\"\n",
       "input: \"1243\"\n",
       "input: \"1233\"\n",
       "output: \"1247\"\n",
       "name: \"Where_712\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1222\"\n",
       "input: \"1247\"\n",
       "output: \"1248\"\n",
       "name: \"Expand_713\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1248\"\n",
       "output: \"1249\"\n",
       "name: \"Unsqueeze_714\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1233\"\n",
       "output: \"1250\"\n",
       "name: \"Shape_715\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1250\"\n",
       "output: \"1251\"\n",
       "name: \"ConstantOfShape_716\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1252\"\n",
       "name: \"Constant_717\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1251\"\n",
       "input: \"1252\"\n",
       "output: \"1253\"\n",
       "name: \"Mul_718\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1233\"\n",
       "input: \"1253\"\n",
       "output: \"1254\"\n",
       "name: \"Equal_719\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1254\"\n",
       "input: \"1251\"\n",
       "input: \"1233\"\n",
       "output: \"1255\"\n",
       "name: \"Where_720\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1224\"\n",
       "input: \"1255\"\n",
       "output: \"1256\"\n",
       "name: \"Expand_721\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1256\"\n",
       "output: \"1257\"\n",
       "name: \"Unsqueeze_722\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1233\"\n",
       "output: \"1258\"\n",
       "name: \"Shape_723\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1258\"\n",
       "output: \"1259\"\n",
       "name: \"ConstantOfShape_724\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1260\"\n",
       "name: \"Constant_725\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1259\"\n",
       "input: \"1260\"\n",
       "output: \"1261\"\n",
       "name: \"Mul_726\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1233\"\n",
       "input: \"1261\"\n",
       "output: \"1262\"\n",
       "name: \"Equal_727\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1262\"\n",
       "input: \"1259\"\n",
       "input: \"1233\"\n",
       "output: \"1263\"\n",
       "name: \"Where_728\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1226\"\n",
       "input: \"1263\"\n",
       "output: \"1264\"\n",
       "name: \"Expand_729\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1264\"\n",
       "output: \"1265\"\n",
       "name: \"Unsqueeze_730\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1233\"\n",
       "output: \"1266\"\n",
       "name: \"Shape_731\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1266\"\n",
       "output: \"1267\"\n",
       "name: \"ConstantOfShape_732\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1268\"\n",
       "name: \"Constant_733\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1267\"\n",
       "input: \"1268\"\n",
       "output: \"1269\"\n",
       "name: \"Mul_734\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1233\"\n",
       "input: \"1269\"\n",
       "output: \"1270\"\n",
       "name: \"Equal_735\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1270\"\n",
       "input: \"1267\"\n",
       "input: \"1233\"\n",
       "output: \"1271\"\n",
       "name: \"Where_736\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1228\"\n",
       "input: \"1271\"\n",
       "output: \"1272\"\n",
       "name: \"Expand_737\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1272\"\n",
       "output: \"1273\"\n",
       "name: \"Unsqueeze_738\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1241\"\n",
       "input: \"1249\"\n",
       "input: \"1257\"\n",
       "input: \"1265\"\n",
       "input: \"1273\"\n",
       "output: \"1274\"\n",
       "name: \"Concat_739\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: -1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"yolo2\"\n",
       "output: \"1275\"\n",
       "name: \"Shape_740\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1276\"\n",
       "name: \"Constant_741\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1277\"\n",
       "name: \"Constant_742\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\005\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1278\"\n",
       "name: \"Constant_743\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\177\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1275\"\n",
       "input: \"1277\"\n",
       "input: \"1278\"\n",
       "input: \"1276\"\n",
       "output: \"1279\"\n",
       "name: \"Slice_744\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"1233\"\n",
       "input: \"1279\"\n",
       "output: \"1280\"\n",
       "name: \"Concat_745\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1178\"\n",
       "input: \"1280\"\n",
       "output: \"1281\"\n",
       "name: \"Reshape_746\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"yolo2\"\n",
       "input: \"1274\"\n",
       "input: \"1281\"\n",
       "output: \"1282\"\n",
       "name: \"ScatterND_747\"\n",
       "op_type: \"ScatterND\"\n",
       ", output: \"1283\"\n",
       "name: \"Constant_748\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1284\"\n",
       "name: \"Constant_749\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1285\"\n",
       "name: \"Constant_750\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1286\"\n",
       "name: \"Constant_751\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1282\"\n",
       "input: \"1284\"\n",
       "input: \"1285\"\n",
       "input: \"1283\"\n",
       "input: \"1286\"\n",
       "output: \"1287\"\n",
       "name: \"Slice_752\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"1287\"\n",
       "output: \"1288\"\n",
       "name: \"Exp_753\"\n",
       "op_type: \"Exp\"\n",
       ", output: \"1289\"\n",
       "name: \"Constant_754\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    dims: 3\n",
       "    dims: 1\n",
       "    dims: 1\n",
       "    dims: 2\n",
       "    data_type: 1\n",
       "    raw_data: \"\\000\\000@?\\000\\000\\220?\\000\\000\\024@\\000\\000D@\\000\\000P@\\000\\000\\004A\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1288\"\n",
       "input: \"1289\"\n",
       "output: \"1290\"\n",
       "name: \"Mul_755\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"1291\"\n",
       "name: \"Constant_756\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1290\"\n",
       "input: \"1291\"\n",
       "output: \"1292\"\n",
       "name: \"Reshape_757\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1293\"\n",
       "name: \"Constant_758\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1555\"\n",
       "output: \"1295\"\n",
       "name: \"ConstantOfShape_759\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1296\"\n",
       "name: \"Constant_760\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1295\"\n",
       "input: \"1296\"\n",
       "output: \"1297\"\n",
       "name: \"Mul_761\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"1298\"\n",
       "name: \"Constant_762\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1298\"\n",
       "input: \"1297\"\n",
       "output: \"1299\"\n",
       "name: \"Equal_763\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1299\"\n",
       "input: \"1295\"\n",
       "input: \"1293\"\n",
       "output: \"1300\"\n",
       "name: \"Where_764\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1292\"\n",
       "input: \"1300\"\n",
       "output: \"1301\"\n",
       "name: \"Expand_765\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1282\"\n",
       "output: \"1302\"\n",
       "name: \"Shape_766\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1303\"\n",
       "name: \"Constant_767\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1302\"\n",
       "input: \"1303\"\n",
       "output: \"1304\"\n",
       "name: \"Gather_768\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1304\"\n",
       "output: \"1305\"\n",
       "name: \"Cast_769\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1306\"\n",
       "name: \"Constant_770\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1307\"\n",
       "name: \"Constant_771\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1306\"\n",
       "input: \"1305\"\n",
       "input: \"1307\"\n",
       "output: \"1308\"\n",
       "name: \"Range_772\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1282\"\n",
       "output: \"1309\"\n",
       "name: \"Shape_773\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1310\"\n",
       "name: \"Constant_774\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1309\"\n",
       "input: \"1310\"\n",
       "output: \"1311\"\n",
       "name: \"Gather_775\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1311\"\n",
       "output: \"1312\"\n",
       "name: \"Cast_776\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1313\"\n",
       "name: \"Constant_777\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1314\"\n",
       "name: \"Constant_778\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1313\"\n",
       "input: \"1312\"\n",
       "input: \"1314\"\n",
       "output: \"1315\"\n",
       "name: \"Range_779\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1282\"\n",
       "output: \"1316\"\n",
       "name: \"Shape_780\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1317\"\n",
       "name: \"Constant_781\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1316\"\n",
       "input: \"1317\"\n",
       "output: \"1318\"\n",
       "name: \"Gather_782\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1318\"\n",
       "output: \"1319\"\n",
       "name: \"Cast_783\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1320\"\n",
       "name: \"Constant_784\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1321\"\n",
       "name: \"Constant_785\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1320\"\n",
       "input: \"1319\"\n",
       "input: \"1321\"\n",
       "output: \"1322\"\n",
       "name: \"Range_786\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1282\"\n",
       "output: \"1323\"\n",
       "name: \"Shape_787\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1324\"\n",
       "name: \"Constant_788\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1323\"\n",
       "input: \"1324\"\n",
       "output: \"1325\"\n",
       "name: \"Gather_789\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1325\"\n",
       "output: \"1326\"\n",
       "name: \"Cast_790\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1327\"\n",
       "name: \"Constant_791\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1328\"\n",
       "name: \"Constant_792\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1327\"\n",
       "input: \"1326\"\n",
       "input: \"1328\"\n",
       "output: \"1329\"\n",
       "name: \"Range_793\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1282\"\n",
       "output: \"1330\"\n",
       "name: \"Shape_794\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1331\"\n",
       "name: \"Constant_795\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1330\"\n",
       "input: \"1331\"\n",
       "output: \"1332\"\n",
       "name: \"Gather_796\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1332\"\n",
       "output: \"1333\"\n",
       "name: \"Cast_797\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1334\"\n",
       "name: \"Constant_798\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1335\"\n",
       "name: \"Constant_799\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1334\"\n",
       "input: \"1333\"\n",
       "input: \"1335\"\n",
       "output: \"1336\"\n",
       "name: \"Range_800\"\n",
       "op_type: \"Range\"\n",
       ", output: \"1337\"\n",
       "name: \"Constant_801\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1338\"\n",
       "name: \"Constant_802\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1339\"\n",
       "name: \"Constant_803\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1340\"\n",
       "name: \"Constant_804\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1336\"\n",
       "input: \"1338\"\n",
       "input: \"1339\"\n",
       "input: \"1337\"\n",
       "input: \"1340\"\n",
       "output: \"1341\"\n",
       "name: \"Slice_805\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"1342\"\n",
       "name: \"Constant_806\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1308\"\n",
       "input: \"1342\"\n",
       "output: \"1343\"\n",
       "name: \"Reshape_807\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1344\"\n",
       "name: \"Constant_808\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1315\"\n",
       "input: \"1344\"\n",
       "output: \"1345\"\n",
       "name: \"Reshape_809\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1346\"\n",
       "name: \"Constant_810\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 3\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1322\"\n",
       "input: \"1346\"\n",
       "output: \"1347\"\n",
       "name: \"Reshape_811\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1348\"\n",
       "name: \"Constant_812\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 2\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1329\"\n",
       "input: \"1348\"\n",
       "output: \"1349\"\n",
       "name: \"Reshape_813\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1350\"\n",
       "name: \"Constant_814\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1341\"\n",
       "input: \"1350\"\n",
       "output: \"1351\"\n",
       "name: \"Reshape_815\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1343\"\n",
       "input: \"1345\"\n",
       "output: \"1352\"\n",
       "name: \"Add_816\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1352\"\n",
       "input: \"1347\"\n",
       "output: \"1353\"\n",
       "name: \"Add_817\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1353\"\n",
       "input: \"1349\"\n",
       "output: \"1354\"\n",
       "name: \"Add_818\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1354\"\n",
       "input: \"1351\"\n",
       "output: \"1355\"\n",
       "name: \"Add_819\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1355\"\n",
       "output: \"1356\"\n",
       "name: \"Shape_820\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1356\"\n",
       "output: \"1357\"\n",
       "name: \"Shape_821\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1357\"\n",
       "output: \"1358\"\n",
       "name: \"ConstantOfShape_822\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1359\"\n",
       "name: \"Constant_823\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1358\"\n",
       "input: \"1359\"\n",
       "output: \"1360\"\n",
       "name: \"Mul_824\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1356\"\n",
       "input: \"1360\"\n",
       "output: \"1361\"\n",
       "name: \"Equal_825\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1361\"\n",
       "input: \"1358\"\n",
       "input: \"1356\"\n",
       "output: \"1362\"\n",
       "name: \"Where_826\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1343\"\n",
       "input: \"1362\"\n",
       "output: \"1363\"\n",
       "name: \"Expand_827\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1363\"\n",
       "output: \"1364\"\n",
       "name: \"Unsqueeze_828\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1356\"\n",
       "output: \"1365\"\n",
       "name: \"Shape_829\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1365\"\n",
       "output: \"1366\"\n",
       "name: \"ConstantOfShape_830\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1367\"\n",
       "name: \"Constant_831\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1366\"\n",
       "input: \"1367\"\n",
       "output: \"1368\"\n",
       "name: \"Mul_832\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1356\"\n",
       "input: \"1368\"\n",
       "output: \"1369\"\n",
       "name: \"Equal_833\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1369\"\n",
       "input: \"1366\"\n",
       "input: \"1356\"\n",
       "output: \"1370\"\n",
       "name: \"Where_834\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1345\"\n",
       "input: \"1370\"\n",
       "output: \"1371\"\n",
       "name: \"Expand_835\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1371\"\n",
       "output: \"1372\"\n",
       "name: \"Unsqueeze_836\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1356\"\n",
       "output: \"1373\"\n",
       "name: \"Shape_837\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1373\"\n",
       "output: \"1374\"\n",
       "name: \"ConstantOfShape_838\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1375\"\n",
       "name: \"Constant_839\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1374\"\n",
       "input: \"1375\"\n",
       "output: \"1376\"\n",
       "name: \"Mul_840\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1356\"\n",
       "input: \"1376\"\n",
       "output: \"1377\"\n",
       "name: \"Equal_841\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1377\"\n",
       "input: \"1374\"\n",
       "input: \"1356\"\n",
       "output: \"1378\"\n",
       "name: \"Where_842\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1347\"\n",
       "input: \"1378\"\n",
       "output: \"1379\"\n",
       "name: \"Expand_843\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1379\"\n",
       "output: \"1380\"\n",
       "name: \"Unsqueeze_844\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1356\"\n",
       "output: \"1381\"\n",
       "name: \"Shape_845\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1381\"\n",
       "output: \"1382\"\n",
       "name: \"ConstantOfShape_846\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1383\"\n",
       "name: \"Constant_847\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1382\"\n",
       "input: \"1383\"\n",
       "output: \"1384\"\n",
       "name: \"Mul_848\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1356\"\n",
       "input: \"1384\"\n",
       "output: \"1385\"\n",
       "name: \"Equal_849\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1385\"\n",
       "input: \"1382\"\n",
       "input: \"1356\"\n",
       "output: \"1386\"\n",
       "name: \"Where_850\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1349\"\n",
       "input: \"1386\"\n",
       "output: \"1387\"\n",
       "name: \"Expand_851\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1387\"\n",
       "output: \"1388\"\n",
       "name: \"Unsqueeze_852\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1356\"\n",
       "output: \"1389\"\n",
       "name: \"Shape_853\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1389\"\n",
       "output: \"1390\"\n",
       "name: \"ConstantOfShape_854\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1391\"\n",
       "name: \"Constant_855\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1390\"\n",
       "input: \"1391\"\n",
       "output: \"1392\"\n",
       "name: \"Mul_856\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1356\"\n",
       "input: \"1392\"\n",
       "output: \"1393\"\n",
       "name: \"Equal_857\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1393\"\n",
       "input: \"1390\"\n",
       "input: \"1356\"\n",
       "output: \"1394\"\n",
       "name: \"Where_858\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1351\"\n",
       "input: \"1394\"\n",
       "output: \"1395\"\n",
       "name: \"Expand_859\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1395\"\n",
       "output: \"1396\"\n",
       "name: \"Unsqueeze_860\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1364\"\n",
       "input: \"1372\"\n",
       "input: \"1380\"\n",
       "input: \"1388\"\n",
       "input: \"1396\"\n",
       "output: \"1397\"\n",
       "name: \"Concat_861\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: -1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1282\"\n",
       "output: \"1398\"\n",
       "name: \"Shape_862\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1399\"\n",
       "name: \"Constant_863\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1400\"\n",
       "name: \"Constant_864\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\005\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1401\"\n",
       "name: \"Constant_865\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\177\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1398\"\n",
       "input: \"1400\"\n",
       "input: \"1401\"\n",
       "input: \"1399\"\n",
       "output: \"1402\"\n",
       "name: \"Slice_866\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"1356\"\n",
       "input: \"1402\"\n",
       "output: \"1403\"\n",
       "name: \"Concat_867\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1301\"\n",
       "input: \"1403\"\n",
       "output: \"1404\"\n",
       "name: \"Reshape_868\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1282\"\n",
       "input: \"1397\"\n",
       "input: \"1404\"\n",
       "output: \"1405\"\n",
       "name: \"ScatterND_869\"\n",
       "op_type: \"ScatterND\"\n",
       ", output: \"1406\"\n",
       "name: \"Constant_870\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1407\"\n",
       "name: \"Constant_871\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1408\"\n",
       "name: \"Constant_872\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1409\"\n",
       "name: \"Constant_873\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1405\"\n",
       "input: \"1407\"\n",
       "input: \"1408\"\n",
       "input: \"1406\"\n",
       "input: \"1409\"\n",
       "output: \"1410\"\n",
       "name: \"Slice_874\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"1411\"\n",
       "name: \"Constant_875\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 1\n",
       "    raw_data: \"\\000\\000\\200A\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1410\"\n",
       "input: \"1411\"\n",
       "output: \"1412\"\n",
       "name: \"Mul_876\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"1413\"\n",
       "name: \"Constant_877\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1412\"\n",
       "input: \"1413\"\n",
       "output: \"1414\"\n",
       "name: \"Reshape_878\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1415\"\n",
       "name: \"Constant_879\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1556\"\n",
       "output: \"1417\"\n",
       "name: \"ConstantOfShape_880\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1418\"\n",
       "name: \"Constant_881\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1417\"\n",
       "input: \"1418\"\n",
       "output: \"1419\"\n",
       "name: \"Mul_882\"\n",
       "op_type: \"Mul\"\n",
       ", output: \"1420\"\n",
       "name: \"Constant_883\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\\003\\000\\000\\000\\000\\000\\000\\000\\024\\000\\000\\000\\000\\000\\000\\000 \\000\\000\\000\\000\\000\\000\\000\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1420\"\n",
       "input: \"1419\"\n",
       "output: \"1421\"\n",
       "name: \"Equal_884\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1421\"\n",
       "input: \"1417\"\n",
       "input: \"1415\"\n",
       "output: \"1422\"\n",
       "name: \"Where_885\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1414\"\n",
       "input: \"1422\"\n",
       "output: \"1423\"\n",
       "name: \"Expand_886\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1405\"\n",
       "output: \"1424\"\n",
       "name: \"Shape_887\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1425\"\n",
       "name: \"Constant_888\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1424\"\n",
       "input: \"1425\"\n",
       "output: \"1426\"\n",
       "name: \"Gather_889\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1426\"\n",
       "output: \"1427\"\n",
       "name: \"Cast_890\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1428\"\n",
       "name: \"Constant_891\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1429\"\n",
       "name: \"Constant_892\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1428\"\n",
       "input: \"1427\"\n",
       "input: \"1429\"\n",
       "output: \"1430\"\n",
       "name: \"Range_893\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1405\"\n",
       "output: \"1431\"\n",
       "name: \"Shape_894\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1432\"\n",
       "name: \"Constant_895\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1431\"\n",
       "input: \"1432\"\n",
       "output: \"1433\"\n",
       "name: \"Gather_896\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1433\"\n",
       "output: \"1434\"\n",
       "name: \"Cast_897\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1435\"\n",
       "name: \"Constant_898\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1436\"\n",
       "name: \"Constant_899\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1435\"\n",
       "input: \"1434\"\n",
       "input: \"1436\"\n",
       "output: \"1437\"\n",
       "name: \"Range_900\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1405\"\n",
       "output: \"1438\"\n",
       "name: \"Shape_901\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1439\"\n",
       "name: \"Constant_902\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\002\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1438\"\n",
       "input: \"1439\"\n",
       "output: \"1440\"\n",
       "name: \"Gather_903\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1440\"\n",
       "output: \"1441\"\n",
       "name: \"Cast_904\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1442\"\n",
       "name: \"Constant_905\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1443\"\n",
       "name: \"Constant_906\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1442\"\n",
       "input: \"1441\"\n",
       "input: \"1443\"\n",
       "output: \"1444\"\n",
       "name: \"Range_907\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1405\"\n",
       "output: \"1445\"\n",
       "name: \"Shape_908\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1446\"\n",
       "name: \"Constant_909\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\003\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1445\"\n",
       "input: \"1446\"\n",
       "output: \"1447\"\n",
       "name: \"Gather_910\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1447\"\n",
       "output: \"1448\"\n",
       "name: \"Cast_911\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1449\"\n",
       "name: \"Constant_912\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1450\"\n",
       "name: \"Constant_913\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1449\"\n",
       "input: \"1448\"\n",
       "input: \"1450\"\n",
       "output: \"1451\"\n",
       "name: \"Range_914\"\n",
       "op_type: \"Range\"\n",
       ", input: \"1405\"\n",
       "output: \"1452\"\n",
       "name: \"Shape_915\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1453\"\n",
       "name: \"Constant_916\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1452\"\n",
       "input: \"1453\"\n",
       "output: \"1454\"\n",
       "name: \"Gather_917\"\n",
       "op_type: \"Gather\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1454\"\n",
       "output: \"1455\"\n",
       "name: \"Cast_918\"\n",
       "op_type: \"Cast\"\n",
       "attribute {\n",
       "  name: \"to\"\n",
       "  i: 7\n",
       "  type: INT\n",
       "}\n",
       ", output: \"1456\"\n",
       "name: \"Constant_919\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1457\"\n",
       "name: \"Constant_920\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1456\"\n",
       "input: \"1455\"\n",
       "input: \"1457\"\n",
       "output: \"1458\"\n",
       "name: \"Range_921\"\n",
       "op_type: \"Range\"\n",
       ", output: \"1459\"\n",
       "name: \"Constant_922\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1460\"\n",
       "name: \"Constant_923\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1461\"\n",
       "name: \"Constant_924\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\004\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1462\"\n",
       "name: \"Constant_925\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1458\"\n",
       "input: \"1460\"\n",
       "input: \"1461\"\n",
       "input: \"1459\"\n",
       "input: \"1462\"\n",
       "output: \"1463\"\n",
       "name: \"Slice_926\"\n",
       "op_type: \"Slice\"\n",
       ", output: \"1464\"\n",
       "name: \"Constant_927\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 5\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1430\"\n",
       "input: \"1464\"\n",
       "output: \"1465\"\n",
       "name: \"Reshape_928\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1466\"\n",
       "name: \"Constant_929\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 4\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1437\"\n",
       "input: \"1466\"\n",
       "output: \"1467\"\n",
       "name: \"Reshape_930\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1468\"\n",
       "name: \"Constant_931\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 3\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1444\"\n",
       "input: \"1468\"\n",
       "output: \"1469\"\n",
       "name: \"Reshape_932\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1470\"\n",
       "name: \"Constant_933\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 2\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1451\"\n",
       "input: \"1470\"\n",
       "output: \"1471\"\n",
       "name: \"Reshape_934\"\n",
       "op_type: \"Reshape\"\n",
       ", output: \"1472\"\n",
       "name: \"Constant_935\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1463\"\n",
       "input: \"1472\"\n",
       "output: \"1473\"\n",
       "name: \"Reshape_936\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1465\"\n",
       "input: \"1467\"\n",
       "output: \"1474\"\n",
       "name: \"Add_937\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1474\"\n",
       "input: \"1469\"\n",
       "output: \"1475\"\n",
       "name: \"Add_938\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1475\"\n",
       "input: \"1471\"\n",
       "output: \"1476\"\n",
       "name: \"Add_939\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1476\"\n",
       "input: \"1473\"\n",
       "output: \"1477\"\n",
       "name: \"Add_940\"\n",
       "op_type: \"Add\"\n",
       ", input: \"1477\"\n",
       "output: \"1478\"\n",
       "name: \"Shape_941\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1478\"\n",
       "output: \"1479\"\n",
       "name: \"Shape_942\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1479\"\n",
       "output: \"1480\"\n",
       "name: \"ConstantOfShape_943\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1481\"\n",
       "name: \"Constant_944\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1480\"\n",
       "input: \"1481\"\n",
       "output: \"1482\"\n",
       "name: \"Mul_945\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1478\"\n",
       "input: \"1482\"\n",
       "output: \"1483\"\n",
       "name: \"Equal_946\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1483\"\n",
       "input: \"1480\"\n",
       "input: \"1478\"\n",
       "output: \"1484\"\n",
       "name: \"Where_947\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1465\"\n",
       "input: \"1484\"\n",
       "output: \"1485\"\n",
       "name: \"Expand_948\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1485\"\n",
       "output: \"1486\"\n",
       "name: \"Unsqueeze_949\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1478\"\n",
       "output: \"1487\"\n",
       "name: \"Shape_950\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1487\"\n",
       "output: \"1488\"\n",
       "name: \"ConstantOfShape_951\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1489\"\n",
       "name: \"Constant_952\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1488\"\n",
       "input: \"1489\"\n",
       "output: \"1490\"\n",
       "name: \"Mul_953\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1478\"\n",
       "input: \"1490\"\n",
       "output: \"1491\"\n",
       "name: \"Equal_954\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1491\"\n",
       "input: \"1488\"\n",
       "input: \"1478\"\n",
       "output: \"1492\"\n",
       "name: \"Where_955\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1467\"\n",
       "input: \"1492\"\n",
       "output: \"1493\"\n",
       "name: \"Expand_956\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1493\"\n",
       "output: \"1494\"\n",
       "name: \"Unsqueeze_957\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1478\"\n",
       "output: \"1495\"\n",
       "name: \"Shape_958\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1495\"\n",
       "output: \"1496\"\n",
       "name: \"ConstantOfShape_959\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1497\"\n",
       "name: \"Constant_960\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1496\"\n",
       "input: \"1497\"\n",
       "output: \"1498\"\n",
       "name: \"Mul_961\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1478\"\n",
       "input: \"1498\"\n",
       "output: \"1499\"\n",
       "name: \"Equal_962\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1499\"\n",
       "input: \"1496\"\n",
       "input: \"1478\"\n",
       "output: \"1500\"\n",
       "name: \"Where_963\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1469\"\n",
       "input: \"1500\"\n",
       "output: \"1501\"\n",
       "name: \"Expand_964\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1501\"\n",
       "output: \"1502\"\n",
       "name: \"Unsqueeze_965\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1478\"\n",
       "output: \"1503\"\n",
       "name: \"Shape_966\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1503\"\n",
       "output: \"1504\"\n",
       "name: \"ConstantOfShape_967\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1505\"\n",
       "name: \"Constant_968\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1504\"\n",
       "input: \"1505\"\n",
       "output: \"1506\"\n",
       "name: \"Mul_969\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1478\"\n",
       "input: \"1506\"\n",
       "output: \"1507\"\n",
       "name: \"Equal_970\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1507\"\n",
       "input: \"1504\"\n",
       "input: \"1478\"\n",
       "output: \"1508\"\n",
       "name: \"Where_971\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1471\"\n",
       "input: \"1508\"\n",
       "output: \"1509\"\n",
       "name: \"Expand_972\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1509\"\n",
       "output: \"1510\"\n",
       "name: \"Unsqueeze_973\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1478\"\n",
       "output: \"1511\"\n",
       "name: \"Shape_974\"\n",
       "op_type: \"Shape\"\n",
       ", input: \"1511\"\n",
       "output: \"1512\"\n",
       "name: \"ConstantOfShape_975\"\n",
       "op_type: \"ConstantOfShape\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\001\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1513\"\n",
       "name: \"Constant_976\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\377\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1512\"\n",
       "input: \"1513\"\n",
       "output: \"1514\"\n",
       "name: \"Mul_977\"\n",
       "op_type: \"Mul\"\n",
       ", input: \"1478\"\n",
       "input: \"1514\"\n",
       "output: \"1515\"\n",
       "name: \"Equal_978\"\n",
       "op_type: \"Equal\"\n",
       ", input: \"1515\"\n",
       "input: \"1512\"\n",
       "input: \"1478\"\n",
       "output: \"1516\"\n",
       "name: \"Where_979\"\n",
       "op_type: \"Where\"\n",
       ", input: \"1473\"\n",
       "input: \"1516\"\n",
       "output: \"1517\"\n",
       "name: \"Expand_980\"\n",
       "op_type: \"Expand\"\n",
       ", input: \"1517\"\n",
       "output: \"1518\"\n",
       "name: \"Unsqueeze_981\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: -1\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1486\"\n",
       "input: \"1494\"\n",
       "input: \"1502\"\n",
       "input: \"1510\"\n",
       "input: \"1518\"\n",
       "output: \"1519\"\n",
       "name: \"Concat_982\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: -1\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1405\"\n",
       "output: \"1520\"\n",
       "name: \"Shape_983\"\n",
       "op_type: \"Shape\"\n",
       ", output: \"1521\"\n",
       "name: \"Constant_984\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\000\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1522\"\n",
       "name: \"Constant_985\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\005\\000\\000\\000\\000\\000\\000\\000\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", output: \"1523\"\n",
       "name: \"Constant_986\"\n",
       "op_type: \"Constant\"\n",
       "attribute {\n",
       "  name: \"value\"\n",
       "  t {\n",
       "    dims: 1\n",
       "    data_type: 7\n",
       "    raw_data: \"\\377\\377\\377\\377\\377\\377\\377\\177\"\n",
       "  }\n",
       "  type: TENSOR\n",
       "}\n",
       ", input: \"1520\"\n",
       "input: \"1522\"\n",
       "input: \"1523\"\n",
       "input: \"1521\"\n",
       "output: \"1524\"\n",
       "name: \"Slice_987\"\n",
       "op_type: \"Slice\"\n",
       ", input: \"1478\"\n",
       "input: \"1524\"\n",
       "output: \"1525\"\n",
       "name: \"Concat_988\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1423\"\n",
       "input: \"1525\"\n",
       "output: \"1526\"\n",
       "name: \"Reshape_989\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1405\"\n",
       "input: \"1519\"\n",
       "input: \"1526\"\n",
       "output: \"1527\"\n",
       "name: \"ScatterND_990\"\n",
       "op_type: \"ScatterND\"\n",
       ", input: \"1147\"\n",
       "output: \"1530\"\n",
       "name: \"Unsqueeze_991\"\n",
       "op_type: \"Unsqueeze\"\n",
       "attribute {\n",
       "  name: \"axes\"\n",
       "  ints: 0\n",
       "  type: INTS\n",
       "}\n",
       ", input: \"1530\"\n",
       "input: \"1557\"\n",
       "input: \"1558\"\n",
       "output: \"1533\"\n",
       "name: \"Concat_992\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 0\n",
       "  type: INT\n",
       "}\n",
       ", input: \"1527\"\n",
       "input: \"1533\"\n",
       "output: \"1534\"\n",
       "name: \"Reshape_993\"\n",
       "op_type: \"Reshape\"\n",
       ", input: \"1119\"\n",
       "input: \"1534\"\n",
       "output: \"classes\"\n",
       "name: \"Concat_994\"\n",
       "op_type: \"Concat\"\n",
       "attribute {\n",
       "  name: \"axis\"\n",
       "  i: 1\n",
       "  type: INT\n",
       "}\n",
       "]"
      ]
     },
     "metadata": {},
     "execution_count": 14
    }
   ],
   "source": [
    "onnx_model.graph.node"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "import onnxruntime\n",
    "\n",
    "ort_session = onnxruntime.InferenceSession(\"weights/yolo-fastest.onnx\")\n",
    "\n",
    "def to_numpy(tensor):\n",
    "    return tensor.detach().cpu().numpy() if tensor.requires_grad else tensor.cpu().numpy()\n",
    "\n",
    "# compute ONNX Runtime output prediction\n",
    "x_array = to_numpy(x)\n",
    "ort_inputs = {ort_session.get_inputs()[0].name: x_array}\n",
    "ort_outs = ort_session.run(None, ort_inputs)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "torch.Size([1, 2400, 7])"
      ]
     },
     "metadata": {},
     "execution_count": 29
    }
   ],
   "source": [
    "torch_out[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "output_type": "error",
     "ename": "AssertionError",
     "evalue": "\nNot equal to tolerance rtol=0.001, atol=1e-05\n\nMismatched elements: 7197 / 16800 (42.8%)\nMax absolute difference:      19.835\nMax relative difference:   4.092e+05\n x: array([[[     25.479,       21.32,      57.021, ...,  2.7384e-06,    0.013062,     0.99144],\n        [     48.863,      23.147,      103.85, ...,  3.9784e-05,   0.0075577,     0.99291],\n        [     84.161,          24,      153.27, ...,  4.0627e-05,   0.0075448,     0.99217],...\n y: array([[[     25.479,       21.32,      57.021, ...,     -12.808,     -4.3249,      4.7517],\n        [     48.863,      23.147,      103.85, ...,     -10.132,     -4.8776,      4.9417],\n        [     84.161,          24,      153.27, ...,     -10.111,     -4.8793,      4.8415],...",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-30-29f2a1871e2b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# compare ONNX Runtime and PyTorch results\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mtorch_out_array\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mto_numpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch_out\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mto_numpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch_out\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mto_numpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch_out\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtesting\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0massert_allclose\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch_out_array\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mort_outs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrtol\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1e-03\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0matol\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1e-05\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Exported model has been tested with ONNXRuntime, and the result looks good!\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/numpy/testing/_private/utils.py\u001b[0m in \u001b[0;36massert_allclose\u001b[0;34m(actual, desired, rtol, atol, equal_nan, err_msg, verbose)\u001b[0m\n\u001b[1;32m   1526\u001b[0m     \u001b[0mheader\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m'Not equal to tolerance rtol=%g, atol=%g'\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mrtol\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0matol\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1527\u001b[0m     assert_array_compare(compare, actual, desired, err_msg=str(err_msg),\n\u001b[0;32m-> 1528\u001b[0;31m                          verbose=verbose, header=header, equal_nan=equal_nan)\n\u001b[0m\u001b[1;32m   1529\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1530\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/numpy/testing/_private/utils.py\u001b[0m in \u001b[0;36massert_array_compare\u001b[0;34m(comparison, x, y, err_msg, verbose, header, precision, equal_nan, equal_inf)\u001b[0m\n\u001b[1;32m    838\u001b[0m                                 \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mverbose\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mheader\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mheader\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    839\u001b[0m                                 names=('x', 'y'), precision=precision)\n\u001b[0;32m--> 840\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0mAssertionError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmsg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    841\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    842\u001b[0m         \u001b[0;32mimport\u001b[0m \u001b[0mtraceback\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAssertionError\u001b[0m: \nNot equal to tolerance rtol=0.001, atol=1e-05\n\nMismatched elements: 7197 / 16800 (42.8%)\nMax absolute difference:      19.835\nMax relative difference:   4.092e+05\n x: array([[[     25.479,       21.32,      57.021, ...,  2.7384e-06,    0.013062,     0.99144],\n        [     48.863,      23.147,      103.85, ...,  3.9784e-05,   0.0075577,     0.99291],\n        [     84.161,          24,      153.27, ...,  4.0627e-05,   0.0075448,     0.99217],...\n y: array([[[     25.479,       21.32,      57.021, ...,     -12.808,     -4.3249,      4.7517],\n        [     48.863,      23.147,      103.85, ...,     -10.132,     -4.8776,      4.9417],\n        [     84.161,          24,      153.27, ...,     -10.111,     -4.8793,      4.8415],..."
     ]
    }
   ],
   "source": [
    "# compare ONNX Runtime and PyTorch results\n",
    "torch_out_array = [to_numpy(torch_out[0]), to_numpy(torch_out[1][0]),to_numpy(torch_out[1][1])]\n",
    "np.testing.assert_allclose(torch_out_array[0], ort_outs[0], rtol=1e-03, atol=1e-05)\n",
    "\n",
    "print(\"Exported model has been tested with ONNXRuntime, and the result looks good!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Exported model has been tested with ONNXRuntime, and the result looks good!\n"
     ]
    }
   ],
   "source": [
    "np.testing.assert_allclose(torch_out_array[1], ort_outs[1], rtol=1e-03, atol=1e-05)\n",
    "np.testing.assert_allclose(torch_out_array[2], ort_outs[2], rtol=1e-03, atol=1e-05)\n",
    "\n",
    "print(\"Exported model has been tested with ONNXRuntime, and the result looks good!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "from onnxruntime.quantization import quantize_dynamic, QuantType\n",
    "\n",
    "model_fp32 = 'weights/yolo-fastest.onnx'\n",
    "model_quant = 'weights/yolo-fastest-int8.onnx'\n",
    "quantized_model = quantize_dynamic(model_fp32, model_quant, weight_type=QuantType.QUInt8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "from onnxruntime.quantization import quantize_qat, QuantType\n",
    "\n",
    "model_fp32 = 'weights/yolo-fastest.onnx'\n",
    "model_quant = 'weights/yolo-fastest-int8.onnx'\n",
    "quantized_model = quantize_qat(model_fp32, model_quant)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_numpy(tensor):\n",
    "    return tensor.detach().cpu().numpy() if tensor.requires_grad else tensor.cpu().numpy()\n",
    "\n",
    "batch_size = 1\n",
    "# Input to the model\n",
    "x = torch.randn(batch_size, 3, 320, 512, requires_grad=True)\n",
    "torch_out = model(x)\n",
    "torch_out_array = [to_numpy(torch_out[0]), to_numpy(torch_out[1][0]),to_numpy(torch_out[1][1])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "output_type": "error",
     "ename": "UnimplementedError",
     "evalue": "The Conv2D op currently does not support grouped convolutions on the CPU. A grouped convolution was attempted to be run because the input depth of 8 does not match the filter input depth of 1 [Op:Conv2D]",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mUnimplementedError\u001b[0m                        Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-24-2a52953e4c7c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mx_array\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtranspose\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_array\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m3\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mkeras_model\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'weights/yolo-fastest.h5'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcompile\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m \u001b[0mkeras_out\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkeras_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_array\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    983\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    984\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0menable_auto_cast_variables\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_compute_dtype_object\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 985\u001b[0;31m           \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcall_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    986\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    987\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_activity_regularizer\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/engine/functional.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, inputs, training, mask)\u001b[0m\n\u001b[1;32m    384\u001b[0m     \"\"\"\n\u001b[1;32m    385\u001b[0m     return self._run_internal_graph(\n\u001b[0;32m--> 386\u001b[0;31m         inputs, training=training, mask=mask)\n\u001b[0m\u001b[1;32m    387\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    388\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mcompute_output_shape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_shape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/engine/functional.py\u001b[0m in \u001b[0;36m_run_internal_graph\u001b[0;34m(self, inputs, training, mask)\u001b[0m\n\u001b[1;32m    506\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    507\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnode\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmap_arguments\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensor_dict\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 508\u001b[0;31m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnode\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlayer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    509\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    510\u001b[0m         \u001b[0;31m# Update tensor_dict.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/engine/base_layer.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    983\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    984\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0menable_auto_cast_variables\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_compute_dtype_object\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 985\u001b[0;31m           \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcall_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    986\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    987\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_activity_regularizer\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/keras/layers/convolutional.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m    245\u001b[0m       \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marray_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_compute_causal_padding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    246\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 247\u001b[0;31m     \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_convolution_op\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkernel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    248\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    249\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_bias\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/util/dispatch.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    199\u001b[0m     \u001b[0;34m\"\"\"Call target, and fall back on dispatchers if there is a TypeError.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    200\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 201\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mtarget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    202\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mTypeError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    203\u001b[0m       \u001b[0;31m# Note: convert_to_eager_tensor currently raises a ValueError, not a\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36mconvolution_v2\u001b[0;34m(input, filters, strides, padding, data_format, dilations, name)\u001b[0m\n\u001b[1;32m   1015\u001b[0m       \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1016\u001b[0m       \u001b[0mdilations\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdilations\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1017\u001b[0;31m       name=name)\n\u001b[0m\u001b[1;32m   1018\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1019\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36mconvolution_internal\u001b[0;34m(input, filters, strides, padding, data_format, dilations, name, call_from_convolution, num_spatial_dims)\u001b[0m\n\u001b[1;32m   1145\u001b[0m           \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1146\u001b[0m           \u001b[0mdilations\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdilations\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1147\u001b[0;31m           name=name)\n\u001b[0m\u001b[1;32m   1148\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1149\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0mchannel_index\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m_conv2d_expanded_batch\u001b[0;34m(input, filters, strides, padding, data_format, dilations, name)\u001b[0m\n\u001b[1;32m   2589\u001b[0m         \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2590\u001b[0m         \u001b[0mdilations\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdilations\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2591\u001b[0;31m         name=name)\n\u001b[0m\u001b[1;32m   2592\u001b[0m   return squeeze_batch_dims(\n\u001b[1;32m   2593\u001b[0m       \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name)\u001b[0m\n\u001b[1;32m    936\u001b[0m       \u001b[0;32mreturn\u001b[0m \u001b[0m_result\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    937\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0m_core\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 938\u001b[0;31m       \u001b[0m_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mraise_from_not_ok_status\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    939\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0m_core\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_FallbackException\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    940\u001b[0m       \u001b[0;32mpass\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/ops.py\u001b[0m in \u001b[0;36mraise_from_not_ok_status\u001b[0;34m(e, name)\u001b[0m\n\u001b[1;32m   6841\u001b[0m   \u001b[0mmessage\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmessage\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m\" name: \"\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;34m\"\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6842\u001b[0m   \u001b[0;31m# pylint: disable=protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 6843\u001b[0;31m   \u001b[0msix\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mraise_from\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_status_to_exception\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcode\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmessage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   6844\u001b[0m   \u001b[0;31m# pylint: enable=protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6845\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/six.py\u001b[0m in \u001b[0;36mraise_from\u001b[0;34m(value, from_value)\u001b[0m\n",
      "\u001b[0;31mUnimplementedError\u001b[0m: The Conv2D op currently does not support grouped convolutions on the CPU. A grouped convolution was attempted to be run because the input depth of 8 does not match the filter input depth of 1 [Op:Conv2D]"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.models import load_model\n",
    "\n",
    "x_array = to_numpy(x)\n",
    "x_array = np.transpose(x_array, [0, 2, 3, 1])\n",
    "keras_model = load_model('weights/yolo-fastest.h5', compile=False)\n",
    "keras_out = keras_model(x_array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "(320, 512, 1, 3)"
      ]
     },
     "metadata": {},
     "execution_count": 23
    }
   ],
   "source": [
    "np.shape(x_array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "Python 3.7.3 64-bit ('anaconda3': virtualenv)",
   "display_name": "Python 3.7.3 64-bit ('anaconda3': virtualenv)",
   "metadata": {
    "interpreter": {
     "hash": "bbd1dd37a092b7f1b47b572bc72841591b000e42b11a51f7bec7c58a01a3fca1"
    }
   }
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3-final"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}