spatial_transformer.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "%matplotlib inline\n",
    "%load_ext tensorboard\n",
    "import matplotlib\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import imgaug.augmenters as iaa\n",
    "import imgaug as ia\n",
    "from datetime import datetime\n",
    "from math import cos, sin, pi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ia.seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
      "11493376/11490434 [==============================] - 1s 0us/step\n"
     ]
    }
   ],
   "source": [
    "\n",
    "mnist = tf.keras.datasets.mnist\n",
    "\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "x_train, x_test = x_train / 255.0, x_test / 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = x_train.reshape(-1, 28, 28, 1)\n",
    "x_test = x_test.reshape(-1, 28, 28, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28, 1)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "H, W,_ = x_train[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Pipeline:\n",
    "# (1) Crop images from each side by 1-16px, do not resize the results\n",
    "#     images back to the input size. Keep them at the cropped size.\n",
    "# (2) Horizontally flip 50% of the images.\n",
    "# (3) Blur images using a gaussian kernel with sigma between 0.0 and 3.0.\n",
    "seq = iaa.Sequential([\n",
    "        # Apply affine transformations to some of the images\n",
    "        # - scale to 80-120% of image height/width (each axis independently)\n",
    "        # - translate by -20 to +20 relative to height/width (per axis)\n",
    "        # - rotate by -45 to +45 degrees\n",
    "        # - shear by -16 to +16 degrees\n",
    "        # - order: use nearest neighbour or bilinear interpolation (fast)\n",
    "        # - mode: use any available mode to fill newly created pixels\n",
    "        #         see API or scikit-image for which modes are available\n",
    "        # - cval: if the mode is constant, then use a random brightness\n",
    "        #         for the newly created pixels (e.g. sometimes black,\n",
    "        #         sometimes white)\n",
    "    iaa.OneOf([\n",
    "    \n",
    "        iaa.Affine(\n",
    "            scale={\"x\": (0.6, 1.1), \"y\": (0.5, 1.1)},\n",
    "            translate_percent={\"x\": (-0.2, 0.2), \"y\": (-0.2, 0.2)},\n",
    "            rotate=(-30, 30),\n",
    "            shear=(-15, 15),\n",
    "            order=[0, 1],\n",
    "            cval=(0),\n",
    "        ),\n",
    "        iaa.Affine(\n",
    "            scale={\"x\": (0.6, 1.1), \"y\": (0.6, 1.1)},\n",
    "            order=[0, 1],\n",
    "            cval=(0),\n",
    "        ),\n",
    "        iaa.Affine(\n",
    "            scale={\"x\": (0.6, 0.8), \"y\": (0.6, 0.8)},\n",
    "            translate_percent={\"x\": (-0.2, 0.2), \"y\": (-0.2, 0.2)},\n",
    "            order=[0, 1],\n",
    "            cval=(0),\n",
    "        ),\n",
    "        iaa.Affine(\n",
    "            rotate=(-60, 60),\n",
    "            #shear=(-30, 30),\n",
    "            order=[0, 1],\n",
    "            cval=(0),\n",
    "        ),\n",
    "        iaa.Affine(\n",
    "            shear=(-40, 40),\n",
    "            order=[0, 1],\n",
    "            cval=(0),\n",
    "        ),\n",
    "\n",
    "    ]\n",
    "    )\n",
    "])\n",
    "x_train_aug = seq(images=x_train)  # done by the library\n",
    "x_test_aug = seq(images=x_test)  # done by the library\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "distorted_mnist = {\"x_train_distorted\":x_train_aug, \"x_test_distorted\": x_test_aug, \"y_train\":y_train, \"y_test\":y_test}\n",
    "pickle.dump(distorted_mnist , open( \"distorted_mnist.p\", \"wb\" ) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "distorted_mnist = pickle.load( open( \"distorted_mnist.p\", \"rb\" ) )\n",
    "x_train_aug = distorted_mnist[\"x_train_distorted\"]\n",
    "x_test_aug = distorted_mnist[\"x_test_distorted\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "ename": "OSError",
     "evalue": "SavedModel file does not exist at: mnist_base_line_model/{saved_model.pbtxt|saved_model.pb}",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mOSError\u001b[0m                                   Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-55-3213a31898cd>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mbaseline_model\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'mnist_base_line_model'\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/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/saving/save.py\u001b[0m in \u001b[0;36mload_model\u001b[0;34m(filepath, custom_objects, compile, options)\u001b[0m\n\u001b[1;32m    204\u001b[0m       \u001b[0mfilepath\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpath_to_string\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilepath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    205\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilepath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msix\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstring_types\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--> 206\u001b[0;31m         \u001b[0mloader_impl\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparse_saved_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilepath\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    207\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0msaved_model_load\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilepath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcompile\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    208\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/tensorflow/python/saved_model/loader_impl.py\u001b[0m in \u001b[0;36mparse_saved_model\u001b[0;34m(export_dir)\u001b[0m\n\u001b[1;32m    112\u001b[0m                   (export_dir,\n\u001b[1;32m    113\u001b[0m                    \u001b[0mconstants\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSAVED_MODEL_FILENAME_PBTXT\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 114\u001b[0;31m                    constants.SAVED_MODEL_FILENAME_PB))\n\u001b[0m\u001b[1;32m    115\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    116\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mOSError\u001b[0m: SavedModel file does not exist at: mnist_base_line_model/{saved_model.pbtxt|saved_model.pb}"
     ]
    }
   ],
   "source": [
    "baseline_model=tf.keras.models.load_model('mnist_base_line_model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# the accuracy is low if we directly use the pre-trained model for the distorted dataset\n",
    "baseline_model.evaluate(x_test_aug, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "from math import ceil\n",
    "def draw_samples(images, images_per_row=5):\n",
    "    num = len(images)\n",
    "    per_row = min(images_per_row, num)\n",
    "    rows = ceil(num /per_row)\n",
    "    fig, axs = plt.subplots(rows, per_row)\n",
    "    count = 0 \n",
    "    for i in range(rows):\n",
    "        \n",
    "        for j in range(images_per_row):\n",
    "            count+=1\n",
    "            if (count > num):\n",
    "                break\n",
    "            if rows == 1:\n",
    "                axs[j+i*per_row].imshow(images[j+i*per_row], cmap='gray')\n",
    "            else:\n",
    "                axs[i,j].imshow(images[j+i*per_row], cmap='gray')\n",
    "            \n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(x_train[:30])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(x_train_aug[:30])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "  #tf.keras.layers.MaxPooling2D((2, 2),),\n",
    "  tf.keras.layers.Conv2D(6, (3,3), input_shape=(H, W,1), padding='valid',activation=\"relu\"),\n",
    "  tf.keras.layers.MaxPooling2D((2, 2)),\n",
    "  tf.keras.layers.Conv2D(16, (3,3),padding='valid',activation=\"relu\"),\n",
    "  tf.keras.layers.MaxPooling2D((2, 2)),\n",
    "  tf.keras.layers.Flatten(),\n",
    "  tf.keras.layers.Dense(120, activation='relu'),\n",
    "  tf.keras.layers.Dense(84, activation='relu'),\n",
    "  tf.keras.layers.Dense(10),\n",
    "    #tf.keras.layers.Softmax()\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Flatten(input_shape=(28, 28)),\n",
    "  tf.keras.layers.Dense(128, activation='relu'),\n",
    "  tf.keras.layers.Dropout(0.2),\n",
    "  tf.keras.layers.Dense(10)\n",
    "])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)\n",
    "model.compile(optimizer='adam',\n",
    "              loss=loss_fn,\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "3750/3750 [==============================] - 20s 5ms/step - loss: 0.6887 - accuracy: 0.7725\n",
      "Epoch 2/5\n",
      "3750/3750 [==============================] - 21s 6ms/step - loss: 0.1684 - accuracy: 0.9464\n",
      "Epoch 3/5\n",
      "3750/3750 [==============================] - 24s 6ms/step - loss: 0.1194 - accuracy: 0.9616\n",
      "Epoch 4/5\n",
      "3750/3750 [==============================] - 22s 6ms/step - loss: 0.0924 - accuracy: 0.9707\n",
      "Epoch 5/5\n",
      "3750/3750 [==============================] - 22s 6ms/step - loss: 0.0782 - accuracy: 0.9737\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7feb855b1cc0>"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# completely retrain the original model with mnist dataset + distorted mnist dataset\n",
    "model.fit(tf.concat([x_train_aug, x_train],0), tf.concat([y_train, y_train],0), epochs=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 2ms/step - loss: 0.1898 - accuracy: 0.9434\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.18978169560432434, 0.9434000253677368]"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# not good for distorted data\n",
    "model.evaluate(x_test_aug, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 2ms/step - loss: 0.0430 - accuracy: 0.9874\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.042976558208465576, 0.9873999953269958]"
      ]
     },
     "execution_count": 187,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# still good for undistorted original data\n",
    "model.evaluate(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "def generate_normalized_homo_meshgrids(inputs):\n",
    "    # for x, y in grid, -1 <=x,y<=1\n",
    "    batch_size = tf.shape(inputs)[0]\n",
    "    _, H, W,_ = inputs.shape\n",
    "    x_range = tf.range(W)\n",
    "    y_range = tf.range(H) \n",
    "    x_mesh, y_mesh = tf.meshgrid(x_range, y_range)\n",
    "    x_mesh = (x_mesh/W-0.5)*2\n",
    "    y_mesh = (y_mesh/H-0.5)*2\n",
    "    y_mesh = tf.reshape(y_mesh, (*y_mesh.shape,1))\n",
    "    x_mesh = tf.reshape(x_mesh, (*x_mesh.shape,1))\n",
    "    ones_mesh = tf.ones_like(x_mesh)\n",
    "    homogeneous_grid = tf.concat([x_mesh, y_mesh, ones_mesh],-1)\n",
    "    homogeneous_grid = tf.reshape(homogeneous_grid, (-1, 3,1))\n",
    "    homogeneous_grid = tf.dtypes.cast(homogeneous_grid, tf.float32)\n",
    "    homogeneous_grid = tf.expand_dims(homogeneous_grid, 0)\n",
    "    return tf.tile(homogeneous_grid, [batch_size, 1,1,1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [],
   "source": [
    "def transform_grids(transformations, grids, inputs):\n",
    "    with tf.name_scope(\"transform_grids\"):\n",
    "        trans_matrices=tf.reshape(transformations, (-1, 2,3))\n",
    "        batch_size = tf.shape(trans_matrices)[0]\n",
    "        gs = tf.squeeze(grids, -1)\n",
    "\n",
    "        reprojected_grids = tf.matmul(trans_matrices, gs, transpose_b=True)\n",
    "        # transform grid range from [-1,1) to the range of [0,1)\n",
    "        reprojected_grids = (tf.linalg.matrix_transpose(reprojected_grids) + 1)*0.5\n",
    "        _, H, W, _ = inputs.shape\n",
    "        reprojected_grids = tf.math.multiply(reprojected_grids, [W, H])\n",
    "\n",
    "        return reprojected_grids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_four_neighbors_from_reprojection(inputs, reprojected_grids):\n",
    "    _, H, W, _ = inputs.shape\n",
    "    \n",
    "    x, y = tf.split(reprojected_grids, 2, axis=-1)\n",
    "\n",
    "    x1 = tf.floor(x)\n",
    "    x1 = tf.dtypes.cast(x1, tf.int32)\n",
    "\n",
    "    x2 = x1 + tf.constant(1) \n",
    "\n",
    "    y1 = tf.floor(y)\n",
    "    y1 = tf.dtypes.cast(y1, tf.int32)\n",
    "    y2 = y1 + tf.constant(1) \n",
    "    \n",
    "    y_max = tf.constant(H - 1, dtype=tf.int32)\n",
    "    x_max = tf.constant(W - 1, dtype=tf.int32)\n",
    "    zero = tf.zeros([1], dtype=tf.int32)\n",
    "\n",
    "    x1_safe = tf.clip_by_value(x1, zero, x_max)\n",
    "    y1_safe = tf.clip_by_value(y1, zero, y_max)\n",
    "    x2_safe = tf.clip_by_value(x2, zero, x_max)\n",
    "    y2_safe = tf.clip_by_value(y2, zero, y_max)\n",
    "    return x1_safe, y1_safe, x2_safe, y2_safe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bilinear_sample(inputs, reprojected_grids):\n",
    "    x1, y1, x2, y2 = generate_four_neighbors_from_reprojection(inputs, reprojected_grids)\n",
    "    x1y1 = tf.concat([y1,x1],-1)\n",
    "    x1y2 = tf.concat([y2,x1],-1)\n",
    "    x2y1 = tf.concat([y1,x2],-1)\n",
    "    x2y2 = tf.concat([y2,x2],-1)\n",
    "\n",
    "    pixel_x1y1 = tf.gather_nd(inputs, x1y1, batch_dims=1)\n",
    "    pixel_x1y2 = tf.gather_nd(inputs, x1y2, batch_dims=1)\n",
    "    pixel_x2y1 = tf.gather_nd(inputs, x2y1, batch_dims=1)\n",
    "    pixel_x2y2 = tf.gather_nd(inputs, x2y2, batch_dims=1)\n",
    "    x, y = tf.split(reprojected_grids, 2, axis=-1)\n",
    "    wx = tf.concat([tf.dtypes.cast(x2, tf.float32) - x, x -tf.dtypes.cast(x1, tf.float32)],-1)\n",
    "    wx = tf.expand_dims(wx, -2)\n",
    "    wy = tf.concat([tf.dtypes.cast(y2, tf.float32) - y, y - tf.dtypes.cast(y1, tf.float32)],-1)\n",
    "    wy = tf.expand_dims(wy, -1)\n",
    "    Q = tf.concat([pixel_x1y1, pixel_x1y2, pixel_x2y1, pixel_x2y2], -1)\n",
    "    Q_shape = tf.shape(Q)\n",
    "    Q = tf.reshape(Q, (Q_shape[0], Q_shape[1],2,2))\n",
    "    Q = tf.cast(Q, tf.float32)\n",
    "\n",
    "    r = wx@Q@wy\n",
    "    _, H, W, channels = inputs.shape\n",
    "\n",
    "    r = tf.reshape(r, (-1,H,W,1))\n",
    "    return r\n",
    "\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(3, 6), dtype=float32, numpy=\n",
       "array([[ 1.        ,  0.        ,  0.        ,  0.        ,  0.3       ,\n",
       "         0.35      ],\n",
       "       [ 0.70710677, -0.70710677,  0.        ,  0.70710677,  0.70710677,\n",
       "         0.        ],\n",
       "       [ 1.        ,  0.        ,  0.1       ,  0.        ,  1.        ,\n",
       "         0.1       ]], dtype=float32)>"
      ]
     },
     "execution_count": 192,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs = x_train_aug[:3]\n",
    "transformations = tf.constant([[1,0,0,0,0.3,0.35],[cos(pi/4),-sin(pi/4),0,sin(pi/4),cos(pi/4),0],[1,0,0.1,0,1,0.1]],dtype=tf.float32)\n",
    "transformations\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "grids = generate_normalized_homo_meshgrids(inputs)\n",
    "#print(grids)\n",
    "reprojected_grids = transform_grids(transformations, grids, inputs)\n",
    "#print(reprojected_grids)\n",
    "results = bilinear_sample(inputs, reprojected_grids)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "def spatial_transform_input(inputs, transformations):\n",
    "    grids = generate_normalized_homo_meshgrids(inputs)\n",
    "    reprojected_grids = transform_grids(transformations, grids,inputs)\n",
    "    result = bilinear_sample(inputs, reprojected_grids)\n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Input, Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def create_localization_head(inputs):\n",
    "#     x = Conv2D(32, (3,3),padding='valid',activation=\"relu\")(inputs)\n",
    "    #x = Flatten()(inputs)\n",
    "    #x = Dense(128, activation='relu')(x)\n",
    "    #x = Dense(60, activation='relu')(x)  \n",
    "    #x = MaxPooling2D((2, 2))(inputs)\n",
    "    x = Conv2D(14, (5,5),padding='valid',activation=\"relu\")(inputs)\n",
    "    x = MaxPooling2D((2, 2), strides=2)(x)\n",
    "    x = Conv2D(32, (5,5), padding='valid',activation=\"relu\")(x)\n",
    "    x = MaxPooling2D((2, 2),strides=2)(x)\n",
    "    #x = Conv2D(120, (5,5), padding='valid',activation=\"relu\")(x)\n",
    "    x = Flatten()(x)    \n",
    "    x = Dense(120, activation='relu')(x)\n",
    "    x = Dropout(0.2)(x)\n",
    "    x = Dense(84, activation='relu')(x)\n",
    "    #x = Dropout(0.2)(x)\n",
    "    x = Dense(6, activation=\"linear\", kernel_initializer=\"zeros\",\n",
    "              bias_initializer=lambda shape, dtype: tf.constant([1,0,0,0,1,0], dtype=dtype))(x) # 6 elements to describe the transformation\n",
    "    \n",
    "    return tf.keras.Model(inputs, x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_model(inputs):\n",
    "    x = Conv2D(6, (3,3),padding='valid',activation=\"relu\")(inputs)\n",
    "    x = MaxPooling2D((2, 2))(x)\n",
    "    x = Conv2D(16, (3,3),padding='valid',activation=\"relu\")(x)\n",
    "    x = MaxPooling2D((2, 2))(x)\n",
    "    x = Flatten()(x)\n",
    "    x = Dense(120, activation='relu')(x)\n",
    "    x = Dense(84, activation='relu')(x)\n",
    "    x = Dense(10)(x)\n",
    "    return x\n",
    "\n",
    "    \n",
    "model_baseline = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Flatten(input_shape=(28, 28)),\n",
    "  tf.keras.layers.Dense(128, activation='relu'),\n",
    "  tf.keras.layers.Dropout(0.2),\n",
    "  tf.keras.layers.Dense(10)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = Input(shape=(H,W,1))\n",
    "localization_head= create_localization_head(inputs)\n",
    "x = spatial_transform_input(inputs, localization_head.output)\n",
    "x = run_model(x)\n",
    "#x = model_baseline(x)\n",
    "st= tf.keras.Model(inputs, x)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<KerasTensor: shape=(None, 6) dtype=float32 (created by layer 'model_8')>"
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "localization_head(inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_9\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_5 (InputLayer)            [(None, 28, 28, 1)]  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_18 (Conv2D)              (None, 24, 24, 14)   364         input_5[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_18 (MaxPooling2D) (None, 12, 12, 14)   0           conv2d_18[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_19 (Conv2D)              (None, 8, 8, 32)     11232       max_pooling2d_18[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_19 (MaxPooling2D) (None, 4, 4, 32)     0           conv2d_19[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "flatten_14 (Flatten)            (None, 512)          0           max_pooling2d_19[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "dense_37 (Dense)                (None, 120)          61560       flatten_14[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dropout_9 (Dropout)             (None, 120)          0           dense_37[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.shape_12 (TFOpLamb (4,)                 0           input_5[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_38 (Dense)                (None, 84)           10164       dropout_9[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "tf.__operators__.getitem_16 (Sl ()                   0           tf.compat.v1.shape_12[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "dense_39 (Dense)                (None, 6)            510         dense_38[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "tf.tile_4 (TFOpLambda)          (None, 784, 3, 1)    0           tf.__operators__.getitem_16[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "tf.reshape_12 (TFOpLambda)      (None, 2, 3)         0           dense_39[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.squeeze_4 (TFOpLam (None, 784, 3)       0           tf.tile_4[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "tf.linalg.matmul_12 (TFOpLambda (None, 2, 784)       0           tf.reshape_12[0][0]              \n",
      "                                                                 tf.compat.v1.squeeze_4[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "tf.linalg.matrix_transpose_4 (T (None, 784, 2)       0           tf.linalg.matmul_12[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.__operators__.add_12 (TFOpLa (None, 784, 2)       0           tf.linalg.matrix_transpose_4[0][0\n",
      "__________________________________________________________________________________________________\n",
      "tf.math.multiply_8 (TFOpLambda) (None, 784, 2)       0           tf.__operators__.add_12[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "tf.math.multiply_9 (TFOpLambda) (None, 784, 2)       0           tf.math.multiply_8[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "tf.split_8 (TFOpLambda)         [(None, 784, 1), (No 0           tf.math.multiply_9[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "tf.math.floor_8 (TFOpLambda)    (None, 784, 1)       0           tf.split_8[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.math.floor_9 (TFOpLambda)    (None, 784, 1)       0           tf.split_8[0][1]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_28 (TFOpLambda)         (None, 784, 1)       0           tf.math.floor_8[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_29 (TFOpLambda)         (None, 784, 1)       0           tf.math.floor_9[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "tf.__operators__.add_13 (TFOpLa (None, 784, 1)       0           tf.cast_28[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.__operators__.add_14 (TFOpLa (None, 784, 1)       0           tf.cast_29[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.clip_by_value_18 (TFOpLambda (None, 784, 1)       0           tf.__operators__.add_13[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "tf.clip_by_value_16 (TFOpLambda (None, 784, 1)       0           tf.cast_28[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.clip_by_value_17 (TFOpLambda (None, 784, 1)       0           tf.cast_29[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.clip_by_value_19 (TFOpLambda (None, 784, 1)       0           tf.__operators__.add_14[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_28 (TFOpLambda)       (None, 784, 2)       0           tf.clip_by_value_17[0][0]        \n",
      "                                                                 tf.clip_by_value_16[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_29 (TFOpLambda)       (None, 784, 2)       0           tf.clip_by_value_19[0][0]        \n",
      "                                                                 tf.clip_by_value_16[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_30 (TFOpLambda)       (None, 784, 2)       0           tf.clip_by_value_17[0][0]        \n",
      "                                                                 tf.clip_by_value_18[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_31 (TFOpLambda)       (None, 784, 2)       0           tf.clip_by_value_19[0][0]        \n",
      "                                                                 tf.clip_by_value_18[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.gather_nd_16 (TFOp (None, 784, 1)       0           input_5[0][0]                    \n",
      "                                                                 tf.concat_28[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.gather_nd_17 (TFOp (None, 784, 1)       0           input_5[0][0]                    \n",
      "                                                                 tf.concat_29[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.gather_nd_18 (TFOp (None, 784, 1)       0           input_5[0][0]                    \n",
      "                                                                 tf.concat_30[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.gather_nd_19 (TFOp (None, 784, 1)       0           input_5[0][0]                    \n",
      "                                                                 tf.concat_31[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_34 (TFOpLambda)       (None, 784, 4)       0           tf.compat.v1.gather_nd_16[0][0]  \n",
      "                                                                 tf.compat.v1.gather_nd_17[0][0]  \n",
      "                                                                 tf.compat.v1.gather_nd_18[0][0]  \n",
      "                                                                 tf.compat.v1.gather_nd_19[0][0]  \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_30 (TFOpLambda)         (None, 784, 1)       0           tf.clip_by_value_18[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.split_9 (TFOpLambda)         [(None, 784, 1), (No 0           tf.math.multiply_9[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_31 (TFOpLambda)         (None, 784, 1)       0           tf.clip_by_value_16[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.compat.v1.shape_14 (TFOpLamb (3,)                 0           tf.concat_34[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.math.subtract_16 (TFOpLambda (None, 784, 1)       0           tf.cast_30[0][0]                 \n",
      "                                                                 tf.split_9[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.math.subtract_17 (TFOpLambda (None, 784, 1)       0           tf.split_9[0][0]                 \n",
      "                                                                 tf.cast_31[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.__operators__.getitem_18 (Sl ()                   0           tf.compat.v1.shape_14[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "tf.__operators__.getitem_19 (Sl ()                   0           tf.compat.v1.shape_14[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_32 (TFOpLambda)         (None, 784, 1)       0           tf.clip_by_value_19[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_33 (TFOpLambda)         (None, 784, 1)       0           tf.clip_by_value_17[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_32 (TFOpLambda)       (None, 784, 2)       0           tf.math.subtract_16[0][0]        \n",
      "                                                                 tf.math.subtract_17[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.reshape_13 (TFOpLambda)      (None, None, 2, 2)   0           tf.concat_34[0][0]               \n",
      "                                                                 tf.__operators__.getitem_18[0][0]\n",
      "                                                                 tf.__operators__.getitem_19[0][0]\n",
      "__________________________________________________________________________________________________\n",
      "tf.math.subtract_18 (TFOpLambda (None, 784, 1)       0           tf.cast_32[0][0]                 \n",
      "                                                                 tf.split_9[0][1]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.math.subtract_19 (TFOpLambda (None, 784, 1)       0           tf.split_9[0][1]                 \n",
      "                                                                 tf.cast_33[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.expand_dims_8 (TFOpLambda)   (None, 784, 1, 2)    0           tf.concat_32[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.cast_34 (TFOpLambda)         (None, None, 2, 2)   0           tf.reshape_13[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "tf.concat_33 (TFOpLambda)       (None, 784, 2)       0           tf.math.subtract_18[0][0]        \n",
      "                                                                 tf.math.subtract_19[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "tf.linalg.matmul_13 (TFOpLambda (None, 784, 1, 2)    0           tf.expand_dims_8[0][0]           \n",
      "                                                                 tf.cast_34[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "tf.expand_dims_9 (TFOpLambda)   (None, 784, 2, 1)    0           tf.concat_33[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "tf.linalg.matmul_14 (TFOpLambda (None, 784, 1, 1)    0           tf.linalg.matmul_13[0][0]        \n",
      "                                                                 tf.expand_dims_9[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "tf.reshape_14 (TFOpLambda)      (None, 28, 28, 1)    0           tf.linalg.matmul_14[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_20 (Conv2D)              (None, 26, 26, 6)    60          tf.reshape_14[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_20 (MaxPooling2D) (None, 13, 13, 6)    0           conv2d_20[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_21 (Conv2D)              (None, 11, 11, 16)   880         max_pooling2d_20[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_21 (MaxPooling2D) (None, 5, 5, 16)     0           conv2d_21[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "flatten_15 (Flatten)            (None, 400)          0           max_pooling2d_21[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "dense_40 (Dense)                (None, 120)          48120       flatten_15[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_41 (Dense)                (None, 84)           10164       dense_40[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_42 (Dense)                (None, 10)           850         dense_41[0][0]                   \n",
      "==================================================================================================\n",
      "Total params: 143,904\n",
      "Trainable params: 143,904\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "st.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)\n",
    "st.compile(optimizer=\"adam\",\n",
    "              loss=loss_fn,\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the Keras TensorBoard callback.\n",
    "logdir=\"logs/fit/\" + datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=logdir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "3750/3750 [==============================] - 56s 15ms/step - loss: 0.5007 - accuracy: 0.8447\n",
      "Epoch 2/5\n",
      "3750/3750 [==============================] - 58s 16ms/step - loss: 0.1031 - accuracy: 0.9679\n",
      "Epoch 3/5\n",
      "3750/3750 [==============================] - 58s 16ms/step - loss: 0.0836 - accuracy: 0.9746\n",
      "Epoch 4/5\n",
      "3750/3750 [==============================] - 58s 16ms/step - loss: 0.0745 - accuracy: 0.9771\n",
      "Epoch 5/5\n",
      "3750/3750 [==============================] - 58s 16ms/step - loss: 0.0656 - accuracy: 0.9794\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7feb04644080>"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "st.fit(tf.concat([x_train_aug, x_train],0), tf.concat([y_train, y_train],0), epochs=5,callbacks=[tensorboard_callback])\n",
    "#st.fit(x_train_aug, y_train, epochs=5,callbacks=[tensorboard_callback])\n",
    "\n",
    "#st.fit( x_train,y_train, epochs=5,callbacks=[tensorboard_callback])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: ./spatial_transformer_model/assets\n"
     ]
    }
   ],
   "source": [
    "st.save(\"./spatial_transformer_model\")\n",
    "st.save_weights('./patial_transformer_weights')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%tensorboard --logdir logs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 3s 7ms/step - loss: 0.1141 - accuracy: 0.9646\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.11407934874296188, 0.9646000266075134]"
      ]
     },
     "execution_count": 231,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "st.evaluate(x_test_aug, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 2s 7ms/step - loss: 0.0362 - accuracy: 0.9889\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.03620568662881851, 0.9889000058174133]"
      ]
     },
     "execution_count": 232,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "st.evaluate(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(x_test[:30])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_data = x_test_aug[:30]\n",
    "label = y_test[:30]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "prediction:  [7 2 1 0 4 1 4 9 5 9 0 6 9 0 1 5 9 7 8 4 9 6 6 5 4 0 7 4 0 1]\n",
      "label:       [7 2 1 0 4 1 4 9 5 9 0 6 9 0 1 5 9 7 3 4 9 6 6 5 4 0 7 4 0 1]\n"
     ]
    }
   ],
   "source": [
    "logits = st.predict(test_data)\n",
    "print(\"prediction: \", tf.math.argmax(logits, axis=-1).numpy())\n",
    "print(\"label:      \", label)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [],
   "source": [
    "predicted_transforms = localization_head.predict(test_data)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [],
   "source": [
    "#predicted_transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [],
   "source": [
    "transformed_inputs = spatial_transform_input(test_data, predicted_transforms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(transformed_inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
      "                                 Dload  Upload   Total   Spent    Left  Speed\n",
      "100 3751k  100 3751k    0     0  2385k      0  0:00:01  0:00:01 --:--:-- 2383k\n"
     ]
    }
   ],
   "source": [
    "!curl \"https://raw.githubusercontent.com/daviddao/spatial-transformer-tensorflow/master/data/mnist_sequence1_sample_5distortions5x5.npz\" --output cluttered_mnist/mnist_sequence1_sample_5distortions5x5.npz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000, 28, 28, 1)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "with np.load('cluttered_mnist/mnist_sequence1_sample_5distortions5x5.npz') as cluttered_mnist:\n",
    "    X_train = cluttered_mnist['X_train'].reshape(-1,40,40,1)\n",
    "    Y_train = cluttered_mnist['y_train']\n",
    "    X_test = cluttered_mnist['X_test'].reshape(-1,40,40,1)\n",
    "    Y_test = cluttered_mnist['y_test']\n",
    "    X_train = tf.image.resize(X_train, (28,28))\n",
    "    X_test = tf.image.resize(X_test, (28,28))\n",
    "    print(X_train.shape)\n",
    "    draw_samples(X_train[:30])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam',\n",
    "              loss=loss_fn,\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/6\n",
      "313/313 [==============================] - 1s 4ms/step - loss: 0.5014 - accuracy: 0.8421\n",
      "Epoch 2/6\n",
      "313/313 [==============================] - 1s 4ms/step - loss: 0.2274 - accuracy: 0.9266\n",
      "Epoch 3/6\n",
      "313/313 [==============================] - 1s 4ms/step - loss: 0.1574 - accuracy: 0.9516\n",
      "Epoch 4/6\n",
      "313/313 [==============================] - 1s 4ms/step - loss: 0.1054 - accuracy: 0.9698\n",
      "Epoch 5/6\n",
      "313/313 [==============================] - 1s 4ms/step - loss: 0.0735 - accuracy: 0.9813\n",
      "Epoch 6/6\n",
      "313/313 [==============================] - 1s 4ms/step - loss: 0.0537 - accuracy: 0.9858\n",
      "32/32 [==============================] - 0s 2ms/step - loss: 0.2106 - accuracy: 0.9330\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.21055179834365845, 0.9330000281333923]"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, Y_train, epochs=6)\n",
    "model.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32/32 [==============================] - 0s 2ms/step - loss: 0.2106 - accuracy: 0.9330\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.21055179834365845, 0.9330000281333923]"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/6\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.2248 - accuracy: 0.9307\n",
      "Epoch 2/6\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.1249 - accuracy: 0.9582\n",
      "Epoch 3/6\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0998 - accuracy: 0.9674\n",
      "Epoch 4/6\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0859 - accuracy: 0.9708\n",
      "Epoch 5/6\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0711 - accuracy: 0.9760\n",
      "Epoch 6/6\n",
      "313/313 [==============================] - 5s 15ms/step - loss: 0.0588 - accuracy: 0.9795\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7feb857c5828>"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "st.fit(X_train,Y_train, epochs=6)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32/32 [==============================] - 1s 8ms/step - loss: 0.0885 - accuracy: 0.9710\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.08853834867477417, 0.9710000157356262]"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "st.evaluate(X_test, Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Test_data = X_test[:30]\n",
    "draw_samples(Test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "predicted_transforms = localization_head.predict(Test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "transformed_inputs = spatial_transform_input(Test_data, predicted_transforms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 30 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_samples(transformed_inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}