Image Classifier/src/Classifiers.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TensorFlow version:  2.1.0\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow.keras.layers import Dense, Flatten, Conv2D, Dropout\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras import Model\n",
    "from tensorflow import keras\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib as mp\n",
    "%matplotlib inline\n",
    "import math\n",
    "import numpy as np\n",
    "from datetime import datetime\n",
    "import io\n",
    "import itertools\n",
    "from packaging import version\n",
    "from six.moves import range\n",
    "\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn.metrics\n",
    "\n",
    "print(\"TensorFlow version: \", tf.__version__)\n",
    "assert version.parse(tf.__version__).release[0] >= 2, \\\n",
    "    \"This notebook requires TensorFlow 2.0 or above.\"\n",
    "\n",
    "\n",
    "\"\"\"this is important because i realized the model might not reset\"\"\"\n",
    "tf.keras.backend.clear_session()\n",
    "\n",
    "plt.style.use('seaborn')\n",
    "plt.rcParams['figure.figsize'] = (14.0, 5.0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Table of contents <a name=\"back\"></a>\n",
    "0. [Setup](#setup)\n",
    "1. [Baseline Model](#introduction)\n",
    "2. [Adding Layers](#model2)\n",
    "3. [Adding Convolutions](#model3)\n",
    "3. [Model Comparison](#compare)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SETUP <a name=\"setup\"></a>\n",
    "\n",
    "Random seed for reproduction of results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load and prep of dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist = tf.keras.datasets.mnist\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\n",
    "\n",
    "x2_train = x_train[..., tf.newaxis]\n",
    "x2_test = x_test[..., tf.newaxis]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Batch and shuffle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_ds = tf.data.Dataset.from_tensor_slices((x2_train, y_train)).shuffle(10000).batch(32)\n",
    "test_ds = tf.data.Dataset.from_tensor_slices((x2_test, y_test)).batch(32)\n",
    "\n",
    "class_names = ['0', '1', '2', '3', '4', \n",
    "    '5', '6', '7', '8', '9']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = np.reshape(x_train[0], (-1, 28, 28, 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "add training example to TensorBoard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "!rm -rf logs\n",
    "\n",
    "# Creates a file writer for the log directory. I made it so that different categories contain different stuf\n",
    "logdir = \"logs/sample_data/\"\n",
    "file_writer = tf.summary.create_file_writer(logdir)\n",
    "\n",
    "with file_writer.as_default():\n",
    "    images = np.reshape(x_train[0:25], (-1, 28, 28, 1))\n",
    "    tf.summary.image(\"25 training data examples\", images, max_outputs=25, step=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualization of input images inline + TensorBoard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1008x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_to_image(figure):\n",
    "    buf = io.BytesIO()\n",
    "    plt.savefig(buf, format='png')\n",
    "    plt.close(figure)\n",
    "    buf.seek(0)\n",
    "    # Convert PNG buffer to TF image\n",
    "    image = tf.image.decode_png(buf.getvalue(), channels=4)\n",
    "    # Add the batch dimension\n",
    "    image = tf.expand_dims(image, 0)\n",
    "    return image\n",
    "\n",
    "def image_grid():\n",
    "    figure = plt.figure(figsize=(10,10))\n",
    "    for i in range(25):\n",
    "    # Start next subplot.\n",
    "        plt.subplot(5, 5, i + 1, title=class_names[y_train[i]])\n",
    "        plt.xticks([])\n",
    "        plt.yticks([])\n",
    "        plt.grid(False)\n",
    "        plt.imshow(x_train[i], cmap=plt.cm.binary)\n",
    "    plt.show()\n",
    "    return figure\n",
    "\n",
    "figure = image_grid()\n",
    "with file_writer.as_default():\n",
    "    tf.summary.image(\"Training data\", plot_to_image(figure), step=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Baseline Model <a name=\"introduction\"></a>\n",
    "* Flattening Layer\n",
    "* Dense Layer 1   w/ RELU activation\n",
    "* Dropout Layer 1 w/ 30% dropout rate\n",
    "* Dense Layer 2   w/ RELU activation\n",
    "\n",
    "[BACK](#back)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using model subclassing to build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ModelBaseline(Model):\n",
    "    def __init__(self):\n",
    "        super(ModelBaseline, self).__init__()\n",
    "        self.flat = Flatten()\n",
    "        self.d1 = Dense(45, activation='relu')\n",
    "        self.drop = Dropout(0.3)\n",
    "        self.d2 = Dense(35, activation='relu')\n",
    "    def call(self, x):\n",
    "        x = self.flat(x)\n",
    "        x = self.d1(x)\n",
    "        x = self.drop(x)\n",
    "        return self.d2(x)\n",
    "\n",
    "model=ModelBaseline()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loss function and Optimizer selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# It uses crossentropy between the labels and predictions to compute loss.\n",
    "loss_object = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)\n",
    "optimizer = tf.keras.optimizers.Adam()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Error metric selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loss = tf.keras.metrics.Mean(name='train_loss')\n",
    "train_acc = tf.keras.metrics.SparseCategoricalAccuracy(name='train_accuracy')\n",
    "\n",
    "test_loss = tf.keras.metrics.Mean(name='test_loss')\n",
    "test_acc = tf.keras.metrics.SparseCategoricalAccuracy(name='test_accuracy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model training using gradient tape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def train_step(images, labels):\n",
    "    with tf.GradientTape() as tape:\n",
    "        predictions = model(images, training=True)\n",
    "        loss = loss_object(labels, predictions)\n",
    "    gradients = tape.gradient(loss, model.trainable_variables)\n",
    "    optimizer.apply_gradients(zip(gradients, model.trainable_variables))\n",
    "\n",
    "    train_loss(loss)\n",
    "    train_acc(labels, predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Test function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def test_step(images, labels):\n",
    "    predictions = model(images, training=False)\n",
    "    t_loss = loss_object(labels, predictions)\n",
    "\n",
    "    test_loss(t_loss)\n",
    "    test_acc(labels, predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Add test and train plots to TensorBoard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "!rm -rf logs/gradient_tape\n",
    "\n",
    "train_log_dir = 'logs/gradient_tape/basline_model/train'\n",
    "test_log_dir = 'logs/gradient_tape/baseline_model/test'\n",
    "train_summary_writer = tf.summary.create_file_writer(train_log_dir)\n",
    "test_summary_writer = tf.summary.create_file_writer(test_log_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "EPOCHS = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Layer model_baseline is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "Epoch 1, Loss: 0.6528986692428589, Accuracy: 83.09833526611328, Test Loss: 0.23290494084358215, Test Accuracy: 93.30999755859375\n",
      "Epoch 2, Loss: 0.2939547598361969, Accuracy: 91.40666961669922, Test Loss: 0.17624688148498535, Test Accuracy: 94.09500122070312\n",
      "Epoch 3, Loss: 0.24371175467967987, Accuracy: 92.83833312988281, Test Loss: 0.15563121438026428, Test Accuracy: 94.58000183105469\n",
      "Epoch 4, Loss: 0.22011563181877136, Accuracy: 93.3883285522461, Test Loss: 0.13693475723266602, Test Accuracy: 94.90249633789062\n",
      "Epoch 5, Loss: 0.20518015325069427, Accuracy: 93.79000091552734, Test Loss: 0.13438817858695984, Test Accuracy: 95.11399841308594\n",
      "Epoch 6, Loss: 0.1901978850364685, Accuracy: 94.21666717529297, Test Loss: 0.12158823758363724, Test Accuracy: 95.31666564941406\n",
      "Epoch 7, Loss: 0.18092438578605652, Accuracy: 94.52666473388672, Test Loss: 0.12295806407928467, Test Accuracy: 95.46286010742188\n",
      "Epoch 8, Loss: 0.174808531999588, Accuracy: 94.60166931152344, Test Loss: 0.12022242695093155, Test Accuracy: 95.59000396728516\n",
      "Epoch 9, Loss: 0.16827216744422913, Accuracy: 94.8800048828125, Test Loss: 0.11732736229896545, Test Accuracy: 95.69444274902344\n",
      "Epoch 10, Loss: 0.16347448527812958, Accuracy: 95.0183334350586, Test Loss: 0.11403840035200119, Test Accuracy: 95.79199981689453\n"
     ]
    }
   ],
   "source": [
    "train_loss_results_baseline = []\n",
    "train_accuracy_results_baseline = []\n",
    "test_loss_results_baseline = []\n",
    "test_accuracy_results_baseline = []\n",
    "\n",
    "for epoch in range(EPOCHS):\n",
    "  # Reset the metrics at the start of the next epoch\n",
    "    train_loss.reset_states()\n",
    "    train_acc.reset_states()\n",
    "    test_loss.reset_states()\n",
    "    train_acc.reset_states()\n",
    "\n",
    "    for images, labels in train_ds:\n",
    "        train_step(images, labels)\n",
    "    with train_summary_writer.as_default():\n",
    "        tf.summary.scalar('loss', train_loss.result(), step=epoch)\n",
    "        tf.summary.scalar('accuracy', train_acc.result(), step=epoch)\n",
    "\n",
    "    for test_images, test_labels in test_ds:\n",
    "        test_step(test_images, test_labels)\n",
    "    with test_summary_writer.as_default():\n",
    "        tf.summary.scalar('loss', test_loss.result(), step=epoch)\n",
    "        tf.summary.scalar('accuracy', test_acc.result(), step=epoch)\n",
    "\n",
    "    template = 'Epoch {}, Loss: {}, Accuracy: {}, Test Loss: {}, Test Accuracy: {}'\n",
    "    print(template.format(epoch + 1,\n",
    "                        train_loss.result(),\n",
    "                        train_acc.result() * 100,\n",
    "                        test_loss.result(),\n",
    "                        test_acc.result() * 100))\n",
    "    train_loss_results_baseline.append(train_loss.result())\n",
    "    train_accuracy_results_baseline.append(train_acc.result() * 100)\n",
    "    test_loss_results_baseline.append(test_loss.result())\n",
    "    test_accuracy_results_baseline.append(test_acc.result() * 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], train_loss_results_baseline, label='Training loss', marker='^')\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], test_loss_results_baseline, label='Testing loss', marker='^')\n",
    "ax1.set_title(\"Loss Plot\")\n",
    "ax1.set_xlabel(\"Epoch\")\n",
    "ax1.set_ylabel(\"Loss\")\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], train_accuracy_results_baseline, label='Training accuracy', marker='o')\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], test_accuracy_results_baseline, label='Testing accuracy', marker='o')\n",
    "ax2.set_title(\"Accuracy Plot\")\n",
    "ax2.set_xlabel(\"Epoch\")\n",
    "ax2.set_ylabel(\"%\")\n",
    "ax1.legend()\n",
    "ax2.legend()\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_baseline\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten (Flatten)            multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                multiple                  35325     \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              multiple                  1610      \n",
      "=================================================================\n",
      "Total params: 36,935\n",
      "Trainable params: 36,935\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Adding Layers <a name=\"model2\"></a>\n",
    "* Flattening Layer\n",
    "* Dense Layer 1   w/ RELU activation\n",
    "* Dropout Layer 1 w/ 30% dropout rate\n",
    "* Dense Layer 2   w/ RELU activation\n",
    "* Dense Layer 3   w/ RELU activation\n",
    "* Dense Layer 4   w/ Softmax activation\n",
    "\n",
    "[BACK](#back)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ModelDenser(Model):\n",
    "    def __init__(self):\n",
    "        super(ModelDenser, self).__init__()\n",
    "        self.flat = Flatten()\n",
    "        self.d1 = Dense(45, activation='relu')\n",
    "        self.drop = Dropout(0.3)\n",
    "        self.d2 = Dense(35, activation='relu')\n",
    "        self.d3 = Dense(23, activation='relu')\n",
    "        self.d4 = Dense(10, activation='softmax')\n",
    "    def call(self, x):\n",
    "        x = self.flat(x)\n",
    "        x = self.d1(x)\n",
    "        x = self.drop(x)\n",
    "        x = self.d2(x)\n",
    "        x = self.d3(x)\n",
    "        return self.d4(x)\n",
    "\n",
    "model=ModelDenser()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_object = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)\n",
    "optimizer = tf.keras.optimizers.Adam()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loss = tf.keras.metrics.Mean(name='train_loss')\n",
    "train_acc = tf.keras.metrics.SparseCategoricalAccuracy(name='train_accuracy')\n",
    "\n",
    "test_loss = tf.keras.metrics.Mean(name='test_loss')\n",
    "test_acc = tf.keras.metrics.SparseCategoricalAccuracy(name='test_accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def train_step(images, labels):\n",
    "    with tf.GradientTape() as tape:\n",
    "        predictions = model(images, training=True)\n",
    "        loss = loss_object(labels, predictions)\n",
    "    gradients = tape.gradient(loss, model.trainable_variables)\n",
    "    optimizer.apply_gradients(zip(gradients, model.trainable_variables))\n",
    "\n",
    "    train_loss(loss)\n",
    "    train_acc(labels, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def test_step(images, labels):\n",
    "    predictions = model(images, training=False)\n",
    "    t_loss = loss_object(labels, predictions)\n",
    "\n",
    "    test_loss(t_loss)\n",
    "    test_acc(labels, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_log_dir = 'logs/gradient_tape/basline_dense/train'\n",
    "test_log_dir = 'logs/gradient_tape/baseline_dense/test'\n",
    "train_summary_writer = tf.summary.create_file_writer(train_log_dir)\n",
    "test_summary_writer = tf.summary.create_file_writer(test_log_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "EPOCHS = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Layer model_denser is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "Epoch 1, Loss: 1.6707079410552979, Accuracy: 80.461669921875, Test Loss: 1.5454633235931396, Test Accuracy: 92.06999969482422\n",
      "Epoch 2, Loss: 1.5648046731948853, Accuracy: 90.01000213623047, Test Loss: 1.5266311168670654, Test Accuracy: 92.78500366210938\n",
      "Epoch 3, Loss: 1.550891637802124, Accuracy: 91.20166778564453, Test Loss: 1.5223599672317505, Test Accuracy: 93.1933364868164\n",
      "Epoch 4, Loss: 1.5437707901000977, Accuracy: 91.8699951171875, Test Loss: 1.5197466611862183, Test Accuracy: 93.40999603271484\n",
      "Epoch 5, Loss: 1.5408860445022583, Accuracy: 92.12667083740234, Test Loss: 1.5172237157821655, Test Accuracy: 93.61199951171875\n",
      "Epoch 6, Loss: 1.536344289779663, Accuracy: 92.50166320800781, Test Loss: 1.5186830759048462, Test Accuracy: 93.71499633789062\n",
      "Epoch 7, Loss: 1.5339581966400146, Accuracy: 92.7750015258789, Test Loss: 1.5105727910995483, Test Accuracy: 93.89856719970703\n",
      "Epoch 8, Loss: 1.5316457748413086, Accuracy: 93.02166748046875, Test Loss: 1.5124720335006714, Test Accuracy: 94.0250015258789\n",
      "Epoch 9, Loss: 1.5294430255889893, Accuracy: 93.2066650390625, Test Loss: 1.5116446018218994, Test Accuracy: 94.13444519042969\n",
      "Epoch 10, Loss: 1.5289493799209595, Accuracy: 93.23500061035156, Test Loss: 1.5169167518615723, Test Accuracy: 94.16699981689453\n"
     ]
    }
   ],
   "source": [
    "train_loss_results_denser = []\n",
    "train_accuracy_results_denser = []\n",
    "test_loss_results_denser = []\n",
    "test_accuracy_results_denser = []\n",
    "\n",
    "for epoch in range(EPOCHS):\n",
    "  # Reset the metrics at the start of the next epoch\n",
    "    train_loss.reset_states()\n",
    "    train_acc.reset_states()\n",
    "    test_loss.reset_states()\n",
    "    train_acc.reset_states()\n",
    "\n",
    "    for images, labels in train_ds:\n",
    "        train_step(images, labels)\n",
    "    with train_summary_writer.as_default():\n",
    "        tf.summary.scalar('loss', train_loss.result(), step=epoch)\n",
    "        tf.summary.scalar('accuracy', train_acc.result(), step=epoch)\n",
    "\n",
    "    for test_images, test_labels in test_ds:\n",
    "        test_step(test_images, test_labels)\n",
    "    with test_summary_writer.as_default():\n",
    "        tf.summary.scalar('loss', test_loss.result(), step=epoch)\n",
    "        tf.summary.scalar('accuracy', test_acc.result(), step=epoch)\n",
    "\n",
    "    template = 'Epoch {}, Loss: {}, Accuracy: {}, Test Loss: {}, Test Accuracy: {}'\n",
    "    print(template.format(epoch + 1,\n",
    "                        train_loss.result(),\n",
    "                        train_acc.result() * 100,\n",
    "                        test_loss.result(),\n",
    "                        test_acc.result() * 100))\n",
    "    train_loss_results_denser.append(train_loss.result())\n",
    "    train_accuracy_results_denser.append(train_acc.result() * 100)\n",
    "    test_loss_results_denser.append(test_loss.result())\n",
    "    test_accuracy_results_denser.append(test_acc.result() * 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], train_loss_results_denser, label='Training loss', marker='^')\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], test_loss_results_denser, label='Testing loss', marker='^')\n",
    "ax1.set_title(\"Loss Plot\")\n",
    "ax1.set_xlabel(\"Epoch\")\n",
    "ax1.set_ylabel(\"Loss\")\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], train_accuracy_results_denser, label='Training accuracy', marker='o')\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], test_accuracy_results_denser, label='Testing accuracy', marker='o')\n",
    "ax2.set_title(\"Accuracy Plot\")\n",
    "ax2.set_xlabel(\"Epoch\")\n",
    "ax2.set_ylabel(\"%\")\n",
    "ax1.legend()\n",
    "ax2.legend()\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "pycharm": {
     "is_executing": true
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_denser\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten_1 (Flatten)          multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              multiple                  35325     \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              multiple                  1610      \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              multiple                  828       \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              multiple                  240       \n",
      "=================================================================\n",
      "Total params: 38,003\n",
      "Trainable params: 38,003\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Adding Convolutions <a name=\"model3\"></a>\n",
    "* Convolutional Layer w/ RELU activation\n",
    "* Flattening Layer\n",
    "* Dense Layer 1 w/ RELU activation\n",
    "* Dense Layer 2 w/ Linear activation\n",
    "\n",
    "[BACK](#back)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "class ModelConv(Model):\n",
    "    def __init__(self):\n",
    "        super(ModelConv, self).__init__()\n",
    "        self.conv1 = Conv2D(32, 3, activation='relu')\n",
    "        self.flatten = Flatten()\n",
    "        self.d1 = Dense(128, activation='relu')\n",
    "        self.d2 = Dense(10)\n",
    "    def call(self, x):\n",
    "        x = self.conv1(x)\n",
    "        x = self.flatten(x)\n",
    "        x = self.d1(x)\n",
    "        return self.d2(x)\n",
    "\n",
    "model=ModelConv()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "loss_object = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)\n",
    "optimizer = tf.keras.optimizers.Adam()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "train_loss = tf.keras.metrics.Mean(name='train_loss')\n",
    "train_acc = tf.keras.metrics.SparseCategoricalAccuracy(name='train_accuracy')\n",
    "\n",
    "test_loss = tf.keras.metrics.Mean(name='test_loss')\n",
    "test_acc = tf.keras.metrics.SparseCategoricalAccuracy(name='test_accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def train_step(images, labels):\n",
    "    with tf.GradientTape() as tape:\n",
    "        predictions = model(images, training=True)\n",
    "        loss = loss_object(labels, predictions)\n",
    "    gradients = tape.gradient(loss, model.trainable_variables)\n",
    "    optimizer.apply_gradients(zip(gradients, model.trainable_variables))\n",
    "\n",
    "    train_loss(loss)\n",
    "    train_acc(labels, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def test_step(images, labels):\n",
    "    predictions = model(images, training=False)\n",
    "    t_loss = loss_object(labels, predictions)\n",
    "\n",
    "    test_loss(t_loss)\n",
    "    test_acc(labels, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "train_log_dir = 'logs/gradient_tape/convolutional/train'\n",
    "test_log_dir = 'logs/gradient_tape/convolutional/test'\n",
    "train_summary_writer = tf.summary.create_file_writer(train_log_dir)\n",
    "test_summary_writer = tf.summary.create_file_writer(test_log_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "EPOCHS = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Layer model_conv is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "Epoch 1, Loss: 0.1300947517156601, Accuracy: 96.12166595458984, Test Loss: 0.06562890857458115, Test Accuracy: 97.81999969482422\n",
      "Epoch 2, Loss: 0.04162195697426796, Accuracy: 98.70832824707031, Test Loss: 0.05343993380665779, Test Accuracy: 98.0199966430664\n",
      "Epoch 3, Loss: 0.021526223048567772, Accuracy: 99.32167053222656, Test Loss: 0.05912143364548683, Test Accuracy: 98.07999420166016\n",
      "Epoch 4, Loss: 0.012814410962164402, Accuracy: 99.55500030517578, Test Loss: 0.05898956581950188, Test Accuracy: 98.12999725341797\n",
      "Epoch 5, Loss: 0.009771225973963737, Accuracy: 99.68333435058594, Test Loss: 0.0636601448059082, Test Accuracy: 98.19400024414062\n",
      "Epoch 6, Loss: 0.00682164216414094, Accuracy: 99.78500366210938, Test Loss: 0.06574052572250366, Test Accuracy: 98.22999572753906\n",
      "Epoch 7, Loss: 0.005032407585531473, Accuracy: 99.81500244140625, Test Loss: 0.0768989697098732, Test Accuracy: 98.23285675048828\n",
      "Epoch 8, Loss: 0.0049448697827756405, Accuracy: 99.84500122070312, Test Loss: 0.0748993456363678, Test Accuracy: 98.25875091552734\n",
      "Epoch 9, Loss: 0.004172462038695812, Accuracy: 99.875, Test Loss: 0.07875212281942368, Test Accuracy: 98.27999877929688\n",
      "Epoch 10, Loss: 0.003325331024825573, Accuracy: 99.88500213623047, Test Loss: 0.07143338769674301, Test Accuracy: 98.30400085449219\n"
     ]
    }
   ],
   "source": [
    "train_loss_results_conv = []\n",
    "train_accuracy_results_conv = []\n",
    "test_loss_results_conv = []\n",
    "test_accuracy_results_conv = []\n",
    "\n",
    "for epoch in range(EPOCHS):\n",
    "  # Reset the metrics at the start of the next epoch\n",
    "    train_loss.reset_states()\n",
    "    train_acc.reset_states()\n",
    "    test_loss.reset_states()\n",
    "    train_acc.reset_states()\n",
    "\n",
    "    for images, labels in train_ds:\n",
    "        train_step(images, labels)\n",
    "    with train_summary_writer.as_default():\n",
    "        tf.summary.scalar('loss', train_loss.result(), step=epoch)\n",
    "        tf.summary.scalar('accuracy', train_acc.result(), step=epoch)\n",
    "\n",
    "    for test_images, test_labels in test_ds:\n",
    "        test_step(test_images, test_labels)\n",
    "    with test_summary_writer.as_default():\n",
    "        tf.summary.scalar('loss', test_loss.result(), step=epoch)\n",
    "        tf.summary.scalar('accuracy', test_acc.result(), step=epoch)\n",
    "\n",
    "    template = 'Epoch {}, Loss: {}, Accuracy: {}, Test Loss: {}, Test Accuracy: {}'\n",
    "    print(template.format(epoch + 1,\n",
    "                        train_loss.result(),\n",
    "                        train_acc.result() * 100,\n",
    "                        test_loss.result(),\n",
    "                        test_acc.result() * 100))\n",
    "    train_loss_results_conv.append(train_loss.result())\n",
    "    train_accuracy_results_conv.append(train_acc.result() * 100)\n",
    "    test_loss_results_conv.append(test_loss.result())\n",
    "    test_accuracy_results_conv.append(test_acc.result() * 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], train_loss_results_conv, label='Training loss', marker='^')\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], test_loss_results_conv, label='Testing loss', marker='^')\n",
    "ax1.set_title(\"Loss Plot\")\n",
    "ax1.set_xlabel(\"Epoch\")\n",
    "ax1.set_ylabel(\"Loss\")\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], train_accuracy_results_conv, label='Training accuracy', marker='o')\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], test_accuracy_results_conv, label='Testing accuracy', marker='o')\n",
    "ax2.set_title(\"Accuracy Plot\")\n",
    "ax2.set_xlabel(\"Epoch\")\n",
    "ax2.set_ylabel(\"%\")\n",
    "ax1.legend()\n",
    "ax2.legend()\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "pycharm": {
     "is_executing": true
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_conv\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              multiple                  320       \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              multiple                  2769024   \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              multiple                  1290      \n",
      "=================================================================\n",
      "Total params: 2,770,634\n",
      "Trainable params: 2,770,634\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Comparison <a name=\"compare\"></a>\n",
    "\n",
    "[BACK](#back)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "pycharm": {
     "is_executing": true
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], test_loss_results_baseline, label='Loss Model-1', marker='^')\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], test_loss_results_denser, label='Loss Model-2', marker='^')\n",
    "ax1.plot([x for x in range(1,EPOCHS+1)], test_loss_results_conv, label='Loss Model-3', marker='^')\n",
    "ax1.set_title(\"Testing Loss\")\n",
    "ax1.set_xlabel(\"Epoch\")\n",
    "ax1.set_ylabel(\"Loss\")\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], test_accuracy_results_baseline, label='Accuracy Model-1', marker='o')\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], test_accuracy_results_denser, label='Accuracy Model-2', marker='o')\n",
    "ax2.plot([x for x in range(1,EPOCHS+1)], test_accuracy_results_conv, label='Accuracy Model-3', marker='o')\n",
    "ax2.set_title(\"Testing Accuracy\")\n",
    "ax2.set_xlabel(\"Epoch\")\n",
    "ax2.set_ylabel(\"%\")\n",
    "ax1.legend()\n",
    "ax2.legend()\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Serving TensorBoard on localhost; to expose to the network, use a proxy or pass --bind_all\r\n",
      "TensorBoard 2.1.1 at http://localhost:6006/ (Press CTRL+C to quit)\r\n"
     ]
    }
   ],
   "source": [
    "!tensorboard --logdir logs/"
   ]
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}