mlp_models_trainning.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "gather": {
          "logged": 1664521211631
        }
      },
      "outputs": [],
      "source": [
        "import torchmetrics\n",
        "import neural_net_model_train\n",
        "\n",
        "import utils_data_load\n",
        "\n",
        "train_loader = utils_data_load.train_loader\n",
        "test_loader = utils_data_load.test_loader\n",
        "\n",
        "input_size = (utils_data_load.X_train).shape[1]\n",
        "number_of_classes = len(utils_data_load.labels_dict)\n",
        "\n",
        "channels = 1\n",
        "batch_size= 32\n",
        "\n",
        "max_epochs = 70\n",
        "\n",
        "args_model = {\n",
        "    'number_of_classes':number_of_classes, \n",
        "    'input_shape' : (32, channels, input_size)\n",
        "    }\n",
        "\n",
        "models_trainning = neural_net_model_train.ModelsTrainning(\n",
        "                    train_loader,\n",
        "                    test_loader,\n",
        "                    max_epochs)\n",
        "\n",
        "\n",
        "\n",
        "def computeMean(values, mean_metric=torchmetrics.MeanMetric(nan_strategy=\"warn\")):\n",
        "    mean_metric.update(values)\n",
        "    mean_value = mean_metric.compute()\n",
        "    mean_metric.reset()\n",
        "\n",
        "    return mean_value\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "gather": {
          "logged": 1664521214880
        }
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "# azure setting\n",
        "torch.multiprocessing.set_sharing_strategy('file_system')\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "###  Trainning MLP model obtained from optimization with  evaluation function for accuracy"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "gather": {
          "logged": 1664473227132
        }
      },
      "outputs": [],
      "source": [
        "# the resulting params from optuna optimization:\n",
        "model_parameters = \"\"\"Params:                          \n",
        "                  n_layers: 4                      \n",
        "                  n_units_l0: 148\n",
        "                  dropout_l0: 0.2394077234844161\n",
        "                  n_units_l1: 85\n",
        "                  dropout_l1: 0.1443158058026526\n",
        "                  n_units_l2: 25\n",
        "                  dropout_l2: 0.2711635652530161\n",
        "                  n_units_l3: 94\n",
        "                  dropout_l3: 0.16566341588513586\n",
        "                  learning_rate: 7.481718442497427e-05\n",
        "                  optimizer: RMSprop\n",
        "                Evaluation function: accuracy\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "gather": {
          "logged": 1664521221300
        }
      },
      "outputs": [],
      "source": [
        "%reload_ext autoreload\n",
        "%autoreload 2\n",
        "\n",
        "import torch\n",
        "from torch import nn\n",
        "\n",
        "DEVICE = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "gather": {
          "logged": 1664473227418
        }
      },
      "outputs": [],
      "source": [
        "#the model is made from the best parameters found by optuna \n",
        "#and batchnorm is replaced by groupnorm\n",
        "class MLP_Model(nn.Module):\n",
        "    def __init__(self,args):\n",
        "        super(MLP_Model, self).__init__()\n",
        "        \n",
        "        self.number_of_classes = args[\"number_of_classes\"]\n",
        "        self.input_shape = args[\"input_shape\"]\n",
        "        self.input_features=args[\"input_shape\"][2]\n",
        "        self.batch_size = args[\"input_shape\"][0]\n",
        "        \n",
        "        \n",
        "        self.fltn = nn.Flatten()\n",
        "        \n",
        "        self.fc1 = nn.Linear(self.input_features, 148)\n",
        "        self.gp1 = nn.GroupNorm(num_groups = 37, num_channels= 148)\n",
        "        self.lrelu1 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop1 =   nn.Dropout(p=0.24)\n",
        "\n",
        "        self.fc2 = nn.Linear(148, 85)\n",
        "        self.gp2 = nn.GroupNorm(num_groups=17,num_channels= 85)\n",
        "        self.lrelu2 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop2 =   nn.Dropout(p=0.144)\n",
        "\n",
        "        self.fc3 = nn.Linear(85, 25)\n",
        "        self.gp3 = nn.GroupNorm(num_groups=5,num_channels= 25)\n",
        "        self.lrelu3 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop3 =   nn.Dropout(p=0.27)\n",
        "\n",
        "        self.fc4 = nn.Linear(25, 94)\n",
        "        self.gp4 = nn.GroupNorm(num_groups=2,num_channels= 94)\n",
        "        self.lrelu4 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop4 =   nn.Dropout(p=0.166)\n",
        "\n",
        "        self.fc5 = nn.Linear(94, self.batch_size)\n",
        "        \n",
        "        self.initialize_weights()\n",
        "        \n",
        "    def forward(self, x):\n",
        "        \n",
        "        out = self.fc1(self.fltn(x))\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp1(out)\n",
        "        out = self.lrelu1(out)\n",
        "        out = self.drop1(out)\n",
        "        \n",
        "        out = self.fc2(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp2(out)\n",
        "        out = self.lrelu2(out)\n",
        "        out = self.drop2(out)\n",
        "        \n",
        "        \n",
        "        out = self.fc3(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp3(out)\n",
        "        out = self.lrelu3(out)\n",
        "        out = self.drop3(out)\n",
        "\n",
        "        \n",
        "        out = self.fc4(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp4(out)\n",
        "        out = self.lrelu4(out)\n",
        "        out = self.drop4(out)\n",
        "        \n",
        "        \n",
        "        outputs = self.fc5(out)\n",
        "        assert not torch.isnan(outputs).any()\n",
        "        # print('-----------------------------')\n",
        "        # out = self.logsm(out)\n",
        "        \n",
        "        outputs = outputs.view(self.batch_size, -1)\n",
        "        linear_output = (nn.Linear(outputs.shape[1], self.number_of_classes)).to(DEVICE)\n",
        "        \n",
        "        # apply Linear to get shape of: (batch_size, number_of_classes)\n",
        "        # the Linear is generated in this way to make the output fit the required shape and to let RNN layer be as wide as possible\n",
        "        outputs = linear_output(outputs)\n",
        "        \n",
        "        # reshaped because  has to match the labels target\n",
        "        outputs = outputs.unsqueeze(2)\n",
        "        \n",
        "        return outputs\n",
        "    \n",
        "    \n",
        "    def initialize_weights(self):\n",
        "        for m in self.modules():\n",
        "            if isinstance(m, nn.Linear):\n",
        "                nn.init.kaiming_uniform_(m.weight.data, nonlinearity='leaky_relu')\n",
        "\n",
        "                if m.bias is not None:\n",
        "                    nn.init.constant_(m.bias, 0)\n",
        "\n",
        "            elif isinstance(m, nn.BatchNorm2d):\n",
        "                nn.init.constant_(m.weight.data, 1)\n",
        "                nn.init.constant_(m.bias.data, 0)\n",
        "                "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "gather": {
          "logged": 1664473227580
        }
      },
      "outputs": [],
      "source": [
        "# skip loading best model saved from optimization because has too many parameters \n",
        "def loadModel(mlp_model,optimizer,PATH = \"best_optimized_models/mlp/accuracy/\"+\"mlp_13\"):\n",
        "    checkpoint = torch.load(PATH)\n",
        "    mlp_model_ver1.load_state_dict(checkpoint[\"model\"])\n",
        "    optimizer.load_state_dict(checkpoint[\"optimizer\"])\n",
        "\n",
        "\n",
        "mlp_model_ver1 = MLP_Model(args_model)\n",
        "\n",
        "optimizer = torch.optim.RMSprop(mlp_model_ver1.parameters(), lr=7.481718442497427e-05, alpha=0.99, eps=1e-08, centered=False)   \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "gather": {
          "logged": 1664479465402
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/anaconda/envs/azureml_py38_PT_TF/lib/python3.8/site-packages/torchmetrics/utilities/prints.py:36: UserWarning: Metric `AUROC` will save all targets and predictions in buffer. For large datasets this may lead to large memory footprint.\n",
            "  warnings.warn(*args, **kwargs)\n",
            "100%|██████████| 55932/55932 [04:00<00:00, 232.76it/s]\n",
            "100%|██████████| 55932/55932 [04:06<00:00, 226.97it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.63it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.56it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.68it/s]\n",
            "100%|██████████| 55932/55932 [04:02<00:00, 230.95it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.78it/s]\n",
            "100%|██████████| 55932/55932 [04:04<00:00, 228.55it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.28it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.44it/s]\n",
            "100%|██████████| 55932/55932 [04:02<00:00, 230.22it/s]\n",
            "100%|██████████| 55932/55932 [04:04<00:00, 228.66it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 230.06it/s]\n",
            "100%|██████████| 55932/55932 [04:08<00:00, 225.47it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 230.11it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.82it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.42it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.35it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.61it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 230.04it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.84it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 230.05it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.63it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.59it/s]\n",
            "100%|██████████| 55932/55932 [04:03<00:00, 229.97it/s]\n"
          ]
        }
      ],
      "source": [
        "models_trainning.setParameters(\n",
        "            net_model=mlp_model_ver1,\n",
        "            optimizer=optimizer\n",
        "        )\n",
        "\n",
        "accuracies_scores, loss_scores, f1_scores, auroc_scores = models_trainning.trainNetModel()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "gather": {
          "logged": 1664479465527
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Train set: Average loss: 0.0666, Average accuracy: 2.7135%,\n",
            " \t Average f1_score: 0.0666, Average Area Under ROC: 0.5000 \n",
            "\n"
          ]
        }
      ],
      "source": [
        "print( \"\\nTrain set: Average loss: {:.4f}, Average accuracy: {:.4f}%,\\n \\t Average f1_score: {:.4f}, Average Area Under ROC: {:.4f} \\n\".format(\n",
        "                computeMean(loss_scores),\n",
        "                computeMean(accuracies_scores),\n",
        "                computeMean(f1_scores),\n",
        "                computeMean(auroc_scores),\n",
        "            ), flush=True\n",
        "        )\n",
        "\n",
        "checkpoint = models_trainning.getModelCheckpoint()\n",
        "torch.save(checkpoint, \"best_optimized_models/mlp/accuracy/\"+\"mlp_trainned_\"+ str(max_epochs))   \n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "gather": {
          "logged": 1664479467201
        }
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "from utils.helper_plotting import plotEvaluationResults\n",
        "\n",
        "plotEvaluationResults(accuracies_scores,loss_scores,f1_scores,auroc_scores,max_epochs, label=\"MLP_ver1\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "gather": {
          "logged": 1664479526343
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 46/46 [00:00<00:00, 151.45it/s]\n",
            "100%|██████████| 46/46 [00:00<00:00, 156.89it/s]\n",
            "100%|██████████| 46/46 [00:00<00:00, 159.04it/s]\n",
            "100%|██████████| 46/46 [00:00<00:00, 159.64it/s]\n",
            "100%|██████████| 18951/18951 [00:56<00:00, 333.69it/s]\n"
          ]
        }
      ],
      "source": [
        "# models_trainning.testNetModel()\n",
        "\n",
        "loss_score, accuracies_scores, f1_scores , auroc_scores=neural_net_model_train.testNetModel(mlp_model_ver1, test_loader)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "gather": {
          "logged": 1664479526465
        },
        "jupyter": {
          "outputs_hidden": false,
          "source_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Test set: Average loss: 2.7125, Average accuracy: 0.0690%,\n",
            " \t  Average f1_score: 0.1164, Average Area Under ROC: 0.4954 \n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(\n",
        "        \"\\nTest set: Average loss: {:.4f}, Average accuracy: {:.4f}%,\\n \\t  Average f1_score: {:.4f}, Average Area Under ROC: {:.4f} \\n\".format(\n",
        "            loss_score, accuracies_scores, f1_scores, auroc_scores\n",
        "        ), flush=True\n",
        "    )"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "###  Trainning MLP model obtained from optimization with  evaluation function for f1_score"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "gather": {
          "logged": 1664479526582
        }
      },
      "outputs": [],
      "source": [
        "# the resulting params from optuna optimization:\n",
        "model_parameters = \"\"\"Params: \n",
        "                    n_layers: 7\n",
        "                    n_units_l0: 115\n",
        "                    dropout_l0: 0.21188057241438274\n",
        "                    n_units_l1: 151\n",
        "                    dropout_l1: 0.20908383005814535\n",
        "                    n_units_l2: 139\n",
        "                    dropout_l2: 0.17490465365600258\n",
        "                    n_units_l3: 208\n",
        "                    dropout_l3: 0.4351588652635723\n",
        "                    n_units_l4: 34\n",
        "                    dropout_l4: 0.31068252985882033\n",
        "                    n_units_l5: 217\n",
        "                    dropout_l5: 0.4612289778084963\n",
        "                    n_units_l6: 58\n",
        "                    dropout_l6: 0.14792666261716694\n",
        "                    learning_rate: 0.00024636626958405986\n",
        "                    optimizer: Adam\n",
        "                  Evaluation function: f1_score\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "gather": {
          "logged": 1664521234581
        }
      },
      "outputs": [],
      "source": [
        "#the model is made from the best parameters found by optuna \n",
        "#and batchnorm is replaced by groupnorm\n",
        "class MLP_ModelVer2(nn.Module):\n",
        "    def __init__(self,args):\n",
        "        super(MLP_ModelVer2, self).__init__()\n",
        "        \n",
        "        self.number_of_classes = args[\"number_of_classes\"]\n",
        "        self.input_shape = args[\"input_shape\"]\n",
        "        self.input_features=args[\"input_shape\"][2]\n",
        "        self.batch_size = args[\"input_shape\"][0]\n",
        "        \n",
        "        self.fltn = nn.Flatten()\n",
        "        \n",
        "        self.fc1 = nn.Linear(self.input_features, 115)\n",
        "        self.gp1 = nn.GroupNorm(num_groups = 23, num_channels= 115)\n",
        "        self.lrelu1 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop1 =   nn.Dropout(p=0.21)\n",
        "\n",
        "        self.fc2 = nn.Linear(115, 151)\n",
        "        self.gp2 = nn.GroupNorm(num_groups=1,num_channels= 151)\n",
        "        self.lrelu2 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop2 =   nn.Dropout(p=0.21)\n",
        "\n",
        "        self.fc3 = nn.Linear(151, 139)\n",
        "        self.gp3 = nn.GroupNorm(num_groups=1,num_channels= 139)\n",
        "        self.lrelu3 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop3 =   nn.Dropout(p=0.175)\n",
        "\n",
        "        self.fc4 = nn.Linear(139, 208)\n",
        "        self.gp4 = nn.GroupNorm(num_groups=26,num_channels= 208)\n",
        "        self.lrelu4 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop4 =   nn.Dropout(p=0.44)\n",
        "        \n",
        "        self.fc5 = nn.Linear(208, 34)\n",
        "        self.gp5 = nn.GroupNorm(num_groups=2,num_channels= 34)\n",
        "        self.lrelu5 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop5 =   nn.Dropout(p=0.31)\n",
        "        \n",
        "        self.fc6 = nn.Linear(34, 217)\n",
        "        self.gp6 = nn.GroupNorm(num_groups=31,num_channels= 217)\n",
        "        self.lrelu6 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop6 =   nn.Dropout(p=0.46)\n",
        "        \n",
        "        self.fc7 = nn.Linear(217, 58)\n",
        "        self.gp7 = nn.GroupNorm(num_groups=2,num_channels= 58)\n",
        "        self.lrelu7 = nn.LeakyReLU(inplace=True)\n",
        "        self.drop7 =   nn.Dropout(p=0.148)\n",
        "\n",
        "        self.fc8 = nn.Linear(58, self.batch_size)\n",
        "        \n",
        "        self.initialize_weights()\n",
        "        \n",
        "    def forward(self, x):\n",
        "        \n",
        "        out = self.fc1(self.fltn(x))\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp1(out)\n",
        "        out = self.lrelu1(out)\n",
        "        out = self.drop1(out)\n",
        "        \n",
        "        out = self.fc2(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp2(out)\n",
        "        out = self.lrelu2(out)\n",
        "        out = self.drop2(out)\n",
        "        \n",
        "        \n",
        "        out = self.fc3(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp3(out)\n",
        "        out = self.lrelu3(out)\n",
        "        out = self.drop3(out)\n",
        "\n",
        "        \n",
        "        out = self.fc4(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp4(out)\n",
        "        out = self.lrelu4(out)\n",
        "        out = self.drop4(out)\n",
        "        \n",
        "        \n",
        "        out = self.fc5(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp5(out)\n",
        "        out = self.lrelu5(out)\n",
        "        out = self.drop5(out)\n",
        "        \n",
        "        \n",
        "        out = self.fc6(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp6(out)\n",
        "        out = self.lrelu6(out)\n",
        "        out = self.drop6(out)\n",
        "        \n",
        "        \n",
        "        out = self.fc7(out)\n",
        "        assert not torch.isnan(out).any()\n",
        "        out = self.gp7(out)\n",
        "        out = self.lrelu7(out)\n",
        "        out = self.drop7(out)\n",
        "        \n",
        "        \n",
        "        outputs = self.fc8(out)\n",
        "        assert not torch.isnan(outputs).any()\n",
        "        # print('-----------------------------')\n",
        "        # out = self.logsm(out)\n",
        "        \n",
        "        outputs = outputs.view(self.batch_size, -1)\n",
        "        linear_output = (nn.Linear(outputs.shape[1], self.number_of_classes)).to(DEVICE)\n",
        "        \n",
        "        # apply Linear to get shape of: (batch_size, number_of_classes)\n",
        "        # the Linear is generated in this way to make the output fit the required shape and to let RNN layer be as wide as possible\n",
        "        outputs = linear_output(outputs)\n",
        "        \n",
        "        # reshaped because  has to match the labels target\n",
        "        outputs = outputs.unsqueeze(2)\n",
        "        \n",
        "        return outputs\n",
        "    \n",
        "    \n",
        "    def initialize_weights(self):\n",
        "        for m in self.modules():\n",
        "            if isinstance(m, nn.Linear):\n",
        "                nn.init.kaiming_uniform_(m.weight.data, nonlinearity='leaky_relu')\n",
        "\n",
        "                if m.bias is not None:\n",
        "                    nn.init.constant_(m.bias, 0)\n",
        "\n",
        "            elif isinstance(m, nn.BatchNorm2d):\n",
        "                nn.init.constant_(m.weight.data, 1)\n",
        "                nn.init.constant_(m.bias.data, 0)\n",
        "                "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "gather": {
          "logged": 1664479526879
        }
      },
      "outputs": [],
      "source": [
        "\n",
        "\n",
        "mlp_model_ver2 = MLP_ModelVer2(args_model)\n",
        "\n",
        "optimizer = torch.optim.RMSprop(mlp_model_ver2.parameters(), lr=0.000246, alpha=0.99, eps=1e-08, centered=False)   "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "gather": {
          "logged": 1664488532200
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/anaconda/envs/azureml_py38_PT_TF/lib/python3.8/site-packages/torchmetrics/utilities/prints.py:36: UserWarning: Metric `AUROC` will save all targets and predictions in buffer. For large datasets this may lead to large memory footprint.\n",
            "  warnings.warn(*args, **kwargs)\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.62it/s]\n",
            "100%|██████████| 55932/55932 [05:53<00:00, 158.43it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.77it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.65it/s]\n",
            "100%|██████████| 55932/55932 [05:53<00:00, 158.42it/s]\n",
            "100%|██████████| 55932/55932 [05:53<00:00, 158.20it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.67it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.47it/s]\n",
            "100%|██████████| 55932/55932 [05:51<00:00, 158.99it/s]\n",
            "100%|██████████| 55932/55932 [05:53<00:00, 158.38it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.52it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.64it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.51it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.63it/s]\n",
            "100%|██████████| 55932/55932 [05:51<00:00, 159.09it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.77it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.54it/s]\n",
            "100%|██████████| 55932/55932 [06:00<00:00, 154.99it/s]\n",
            "100%|██████████| 55932/55932 [05:55<00:00, 157.38it/s]\n",
            "100%|██████████| 55932/55932 [05:54<00:00, 157.93it/s]\n",
            "100%|██████████| 55932/55932 [05:53<00:00, 158.08it/s]\n",
            "100%|██████████| 55932/55932 [05:55<00:00, 157.38it/s]\n",
            "100%|██████████| 55932/55932 [05:53<00:00, 158.30it/s]\n",
            "100%|██████████| 55932/55932 [05:52<00:00, 158.60it/s]\n",
            "100%|██████████| 55932/55932 [05:55<00:00, 157.48it/s]\n"
          ]
        }
      ],
      "source": [
        "models_trainning.setParameters(\n",
        "            net_model=mlp_model_ver2,\n",
        "            optimizer=optimizer\n",
        "        )\n",
        "\n",
        "accuracies_scores, loss_scores, f1_scores, auroc_scores = models_trainning.trainNetModel()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "gather": {
          "logged": 1664488532434
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Train set: Average loss: 0.0667, Average accuracy: 2.7130%,\n",
            " \t Average f1_score: 0.0667, Average Area Under ROC: 0.5001 \n",
            "\n"
          ]
        }
      ],
      "source": [
        "print( \"\\nTrain set: Average loss: {:.4f}, Average accuracy: {:.4f}%,\\n \\t Average f1_score: {:.4f}, Average Area Under ROC: {:.4f} \\n\".format(\n",
        "                computeMean(loss_scores),\n",
        "                computeMean(accuracies_scores),\n",
        "                computeMean(f1_scores),\n",
        "                computeMean(auroc_scores),\n",
        "            ), flush=True\n",
        "        )\n",
        "\n",
        "checkpoint = models_trainning.getModelCheckpoint()\n",
        "torch.save(checkpoint, \"best_optimized_models/mlp/f1_score/\"+\"mlp_trainned_\"+ str(max_epochs))   "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "gather": {
          "logged": 1664488533606
        }
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "from utils.helper_plotting import plotEvaluationResults\n",
        "\n",
        "plotEvaluationResults(accuracies_scores,loss_scores,f1_scores,auroc_scores,max_epochs, label=\"MLP_ver2\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "gather": {
          "logged": 1664488610472
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 18951/18951 [01:14<00:00, 255.70it/s]\n"
          ]
        }
      ],
      "source": [
        "# models_trainning.testNetModel()\n",
        "loss_score, accuracies_scores, f1_scores , auroc_scores=neural_net_model_train.testNetModel(mlp_model_ver2, test_loader)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "gather": {
          "logged": 1664488610614
        },
        "jupyter": {
          "outputs_hidden": false,
          "source_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Test set: Average loss: 2.7136, Average accuracy: 0.0673%,\n",
            " \t  Average f1_score: 0.1133, Average Area Under ROC: 0.5073 \n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(\n",
        "        \"\\nTest set: Average loss: {:.4f}, Average accuracy: {:.4f}%,\\n \\t  Average f1_score: {:.4f}, Average Area Under ROC: {:.4f} \\n\".format(\n",
        "            loss_score, accuracies_scores, f1_scores, auroc_scores\n",
        "        ), flush=True\n",
        "    )"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Train the MLP model version 2 for 70 more epochs"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "gather": {
          "logged": 1664521251300
        }
      },
      "outputs": [],
      "source": [
        "mlp_model_ver2 = MLP_ModelVer2(args_model)\n",
        "optimizer = torch.optim.RMSprop(mlp_model_ver2.parameters(), lr=0.000246, alpha=0.99, eps=1e-08, centered=False)   \n",
        "\n",
        "\n",
        "max_epochs_prev=25\n",
        "checkpoint = torch.load(\"best_optimized_models/mlp/f1_score/\"+\"mlp_trainned_\"+ str(max_epochs_prev))\n",
        "mlp_model_ver2.load_state_dict(checkpoint['model'])\n",
        "optimizer.load_state_dict(checkpoint['optimizer'])\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "gather": {
          "logged": 1664550003144
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/anaconda/envs/azureml_py38_PT_TF/lib/python3.8/site-packages/torchmetrics/utilities/prints.py:36: UserWarning: Metric `AUROC` will save all targets and predictions in buffer. For large datasets this may lead to large memory footprint.\n",
            "  warnings.warn(*args, **kwargs)\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.66it/s]\n",
            "100%|██████████| 55932/55932 [06:37<00:00, 140.62it/s]\n",
            "100%|██████████| 55932/55932 [06:34<00:00, 141.89it/s]\n",
            "100%|██████████| 55932/55932 [06:33<00:00, 142.28it/s]\n",
            "100%|██████████| 55932/55932 [06:31<00:00, 142.80it/s]\n",
            "100%|██████████| 55932/55932 [06:33<00:00, 142.04it/s]\n",
            "100%|██████████| 55932/55932 [06:36<00:00, 141.16it/s]\n",
            "100%|██████████| 55932/55932 [06:37<00:00, 140.77it/s]\n",
            "100%|██████████| 55932/55932 [06:33<00:00, 142.13it/s]\n",
            "100%|██████████| 55932/55932 [06:33<00:00, 142.02it/s]\n",
            "100%|██████████| 55932/55932 [06:36<00:00, 140.99it/s]\n",
            "100%|██████████| 55932/55932 [06:34<00:00, 141.68it/s]\n",
            "100%|██████████| 55932/55932 [06:51<00:00, 135.95it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.51it/s]\n",
            "100%|██████████| 55932/55932 [06:32<00:00, 142.42it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.50it/s]\n",
            "100%|██████████| 55932/55932 [06:43<00:00, 138.54it/s]\n",
            "100%|██████████| 55932/55932 [06:41<00:00, 139.43it/s]\n",
            "100%|██████████| 55932/55932 [06:49<00:00, 136.73it/s]\n",
            "100%|██████████| 55932/55932 [06:44<00:00, 138.13it/s]\n",
            "100%|██████████| 55932/55932 [06:39<00:00, 140.05it/s]\n",
            "100%|██████████| 55932/55932 [06:46<00:00, 137.74it/s]\n",
            "100%|██████████| 55932/55932 [06:41<00:00, 139.45it/s]\n",
            "100%|██████████| 55932/55932 [06:38<00:00, 140.35it/s]\n",
            "100%|██████████| 55932/55932 [06:39<00:00, 140.06it/s]\n",
            "100%|██████████| 55932/55932 [06:41<00:00, 139.20it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.54it/s]\n",
            "100%|██████████| 55932/55932 [06:49<00:00, 136.67it/s]\n",
            "100%|██████████| 55932/55932 [06:51<00:00, 135.91it/s]\n",
            "100%|██████████| 55932/55932 [06:45<00:00, 138.09it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.55it/s]\n",
            "100%|██████████| 55932/55932 [06:49<00:00, 136.42it/s]\n",
            "100%|██████████| 55932/55932 [06:58<00:00, 133.76it/s]\n",
            "100%|██████████| 55932/55932 [06:47<00:00, 137.37it/s]\n",
            "100%|██████████| 55932/55932 [06:50<00:00, 136.30it/s]\n",
            "100%|██████████| 55932/55932 [06:41<00:00, 139.21it/s]\n",
            "100%|██████████| 55932/55932 [06:56<00:00, 134.26it/s]\n",
            "100%|██████████| 55932/55932 [07:03<00:00, 132.20it/s]\n",
            "100%|██████████| 55932/55932 [06:49<00:00, 136.64it/s]\n",
            "100%|██████████| 55932/55932 [06:44<00:00, 138.31it/s]\n",
            "100%|██████████| 55932/55932 [07:04<00:00, 131.89it/s]\n",
            "100%|██████████| 55932/55932 [06:44<00:00, 138.40it/s]\n",
            "100%|██████████| 55932/55932 [06:49<00:00, 136.46it/s]\n",
            "100%|██████████| 55932/55932 [06:50<00:00, 136.18it/s]\n",
            "100%|██████████| 55932/55932 [06:55<00:00, 134.55it/s]\n",
            "100%|██████████| 55932/55932 [06:43<00:00, 138.50it/s]\n",
            "100%|██████████| 55932/55932 [06:37<00:00, 140.83it/s]\n",
            "100%|██████████| 55932/55932 [06:35<00:00, 141.27it/s]\n",
            "100%|██████████| 55932/55932 [06:36<00:00, 140.99it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.54it/s]\n",
            "100%|██████████| 55932/55932 [06:35<00:00, 141.58it/s]\n",
            "100%|██████████| 55932/55932 [06:34<00:00, 141.76it/s]\n",
            "100%|██████████| 55932/55932 [06:55<00:00, 134.61it/s]\n",
            "100%|██████████| 55932/55932 [06:38<00:00, 140.20it/s]\n",
            "100%|██████████| 55932/55932 [06:35<00:00, 141.30it/s]\n",
            "100%|██████████| 55932/55932 [06:36<00:00, 141.22it/s]\n",
            "100%|██████████| 55932/55932 [06:32<00:00, 142.60it/s]\n",
            "100%|██████████| 55932/55932 [06:33<00:00, 142.13it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.57it/s]\n",
            "100%|██████████| 55932/55932 [06:33<00:00, 142.27it/s]\n",
            "100%|██████████| 55932/55932 [06:34<00:00, 141.70it/s]\n",
            "100%|██████████| 55932/55932 [06:37<00:00, 140.64it/s]\n",
            "100%|██████████| 55932/55932 [06:44<00:00, 138.27it/s]\n",
            "100%|██████████| 55932/55932 [06:40<00:00, 139.60it/s]\n",
            "100%|██████████| 55932/55932 [06:34<00:00, 141.94it/s]\n",
            "100%|██████████| 55932/55932 [06:30<00:00, 143.06it/s]\n",
            "100%|██████████| 55932/55932 [06:24<00:00, 145.45it/s]\n",
            "100%|██████████| 55932/55932 [06:23<00:00, 145.70it/s]\n",
            "100%|██████████| 55932/55932 [06:22<00:00, 146.34it/s]\n"
          ]
        }
      ],
      "source": [
        "models_trainning.setParameters(\n",
        "            net_model=mlp_model_ver2,\n",
        "            optimizer=optimizer\n",
        "        )\n",
        "\n",
        "accuracies_scores, loss_scores, f1_scores, auroc_scores = models_trainning.trainNetModel()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "gather": {
          "logged": 1664550003400
        }
      },
      "outputs": [],
      "source": [
        "no_epochs = max_epochs + max_epochs_prev\n",
        "\n",
        "checkpoint = models_trainning.getModelCheckpoint()\n",
        "torch.save(checkpoint, \"best_optimized_models/mlp/f1_score/\"+\"mlp_trainned_\"+ str(no_epochs))   "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "gather": {
          "logged": 1664550006005
        }
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "from utils.helper_plotting import plotEvaluationResults\n",
        "\n",
        "plotEvaluationResults(accuracies_scores,loss_scores,f1_scores,auroc_scores,no_epochs, label=\"MLP_ver2\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "gather": {
          "logged": 1664550094182
        }
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 18951/18951 [01:24<00:00, 225.06it/s]\n"
          ]
        }
      ],
      "source": [
        "# models_trainning.testNetModel()\n",
        "\n",
        "loss_score, accuracies_scores, f1_scores , auroc_scores=neural_net_model_train.testNetModel(mlp_model_ver2, test_loader)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "gather": {
          "logged": 1664550094412
        }
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Test set: Average loss: 2.7135, Average accuracy: 0.0671%,\n",
            " \t  Average f1_score: 0.1134, Average Area Under ROC: 0.5053 \n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(\n",
        "        \"\\nTest set: Average loss: {:.4f}, Average accuracy: {:.4f}%,\\n \\t  Average f1_score: {:.4f}, Average Area Under ROC: {:.4f} \\n\".format(\n",
        "            loss_score, accuracies_scores, f1_scores, auroc_scores\n",
        "        ), flush=True\n",
        "    )"
      ]
    }
  ],
  "metadata": {
    "kernel_info": {
      "name": "python38-azureml-pt-tf"
    },
    "kernelspec": {
      "display_name": "Python 3.10.6 ('pytorch_env')",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.10.6"
    },
    "microsoft": {
      "host": {
        "AzureML": {
          "notebookHasBeenCompleted": true
        }
      }
    },
    "nteract": {
      "version": "nteract-front-end@1.0.0"
    },
    "orig_nbformat": 4,
    "vscode": {
      "interpreter": {
        "hash": "201b6068cbed112f118c29a6393440d4ce4dc02105f31305c61d838eb972bb93"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}