MNIST_CNN Notebook.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow.keras import datasets, layers, models\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Download and prepare the MNIST dataset\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The CIFAR10 dataset contains 60,000 color images in 10 classes, with 6,000 images in each class. The dataset is divided into 50,000 training images and 10,000 testing images. The classes are mutually exclusive and there is no overlap between them."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.mnist.load_data()\n",
    "\n",
    "# Normalize pixel values to be between 0 and 1\n",
    "train_images, test_images = train_images / 255.0, test_images / 255.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Verify the data\n",
    "### To verify that the dataset looks correct, let's plot the first 25 images from the training set and display the class name below each image."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(28, 28)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_names = [0,1,2,3,4,5,6,7,8,9]\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "print(train_images[0].shape)\n",
    "for i in range(25):\n",
    "    plt.subplot(5,5,i+1)\n",
    "    plt.xticks([])\n",
    "    plt.yticks([])\n",
    "    plt.grid(False)\n",
    "    plt.imshow(train_images[i], cmap=plt.cm.binary)\n",
    "    # The CIFAR labels happen to be arrays, \n",
    "    # which is why you need the extra index\n",
    "    plt.xlabel(class_names[train_labels[i]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "real_train_labels = []\n",
    "real_test_labels = []\n",
    "for i, val in enumerate(train_labels):\n",
    "    new_arr = [0 for j in range(10)]\n",
    "    new_arr[train_labels[i]] = 1\n",
    "    real_train_labels.append(new_arr)\n",
    "for i, val in enumerate(test_labels):\n",
    "    new_arr = [0 for j in range(10)]\n",
    "    new_arr[test_labels[i]] = 1\n",
    "    real_test_labels.append(new_arr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = models.Sequential()\n",
    "\n",
    "# TODO: Create our linear MNIST model.\n",
    "# Here are each of the layers:\n",
    "# 1) FLATTEN OUR IMAGE so it's a vector (Why do we need to do this?)\n",
    "# 2) Dense layer with output vector of size 256, relu activation\n",
    "# 3) Dense layer with output vector of size 128 , relu activation\n",
    "# 4) Dense layer with output vector of size 64 , relu activation\n",
    "# 5) Final Dense layer. What should the output size here be? (WHAT ACTIVATION FUNCTION TO FINISH IT OFF??? ASK IF YOU DON\"T KNOW!!!)\n",
    "\n",
    "# HINT: you can add Keras layers using model.add(...). Might wanna lookup tf.keras.layers...\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "cnn_model = models.Sequential()\n",
    "\n",
    "# TODO: Create our CNN model.\n",
    "# Here are each of the layers:\n",
    "# 1) convolution using 32 3x3 kernels, relu activation, with input size (28x28x1)\n",
    "# 2) max pool in groups of size 2x2\n",
    "# 3) convolution using 64 3x3 kernels, relu activation\n",
    "# 4) max pool in groups of size 2x2\n",
    "# 5) conv using 64 3x3 kernels, relu activation\n",
    "# 6) Dense layer with output vector of size 64 (what do you need to do before passing into the dense layer?)\n",
    "# 7) Final Dense layer. What should the output size here be? (WHAT ACTIVATION FUNCTION TO FINISH IT OFF??? ASK IF YOU DON\"T KNOW!!!)\n",
    "\n",
    "# HINT: you can add Keras layers using model.add(...). Might wanna lookup tf.keras.layers...\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "module 'tensorflow._api.v1.keras.losses' has no attribute 'SparseCategoricalCrossentropy'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-14-27a636ea8c2f>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlosses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSparseCategoricalCrossentropy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfrom_logits\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m: module 'tensorflow._api.v1.keras.losses' has no attribute 'SparseCategoricalCrossentropy'"
     ]
    }
   ],
   "source": [
    "tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Display the architecture of the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.summary()\n",
    "cnn_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compile and train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: fill in the parameters for model.compile\n",
    "model.compile(...)\n",
    "cnn_model.compile(...)\n",
    "# TODO: train the model after compiling it. Lookup what function to use and what to pass into that function\n",
    "history = ...\n",
    "cnn_history = ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot model accuracy over time/epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.plot(history.history['accuracy'], label='accuracy')\n",
    "plt.plot(history.history['val_accuracy'], label = 'val_accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.ylim([0.5, 1])\n",
    "plt.legend(loc='lower right')\n",
    "\n",
    "test_loss, test_acc = model.evaluate(test_images,  test_labels, verbose=2)\n",
    "cnn_test_loss, cnn_test_acc = model.evaluate(cnn_test_images,  cnn_test_labels, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"test accuracy: {0}\".format(test_acc))\n",
    "print(\"test accuracy: {0}\".format(cnn_test_acc))"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}