MNIST-Simple.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "e98280a2-2321-4290-9f94-77ee5d39e2f7"
    }
   },
   "source": [
    "# Simple MNIST Model\n",
    "\n",
    "In this notebook we will create a simple model to predicts what letter is written in a 28x28 image based on the [MNIST dataset](http://yann.lecun.com/exdb/mnist/).  This is done via a two fully connected layer Neural Network on a flattened version of the image."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "81f31bb2-0012-405d-acb2-5fc4ceeb463d"
    }
   },
   "source": [
    "## Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "nbpresent": {
     "id": "ea336b2d-4ba4-4054-92ad-bf420b56709a"
    }
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import random\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Download Mnist Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-2-8bf8ae5a5303>:2: read_data_sets (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use alternatives such as official/mnist/dataset.py from tensorflow/models.\n",
      "WARNING:tensorflow:From /usr/local/Caskroom/miniconda/base/envs/tensorflow-v1-test/lib/python3.6/site-packages/tensorflow_core/contrib/learn/python/learn/datasets/mnist.py:260: maybe_download (from tensorflow.contrib.learn.python.learn.datasets.base) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please write your own downloading logic.\n",
      "WARNING:tensorflow:From /usr/local/Caskroom/miniconda/base/envs/tensorflow-v1-test/lib/python3.6/site-packages/tensorflow_core/contrib/learn/python/learn/datasets/mnist.py:262: extract_images (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use tf.data to implement this functionality.\n",
      "Extracting MNIST_data/train-images-idx3-ubyte.gz\n",
      "WARNING:tensorflow:From /usr/local/Caskroom/miniconda/base/envs/tensorflow-v1-test/lib/python3.6/site-packages/tensorflow_core/contrib/learn/python/learn/datasets/mnist.py:267: extract_labels (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use tf.data to implement this functionality.\n",
      "Extracting MNIST_data/train-labels-idx1-ubyte.gz\n",
      "WARNING:tensorflow:From /usr/local/Caskroom/miniconda/base/envs/tensorflow-v1-test/lib/python3.6/site-packages/tensorflow_core/contrib/learn/python/learn/datasets/mnist.py:110: dense_to_one_hot (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use tf.one_hot on tensors.\n",
      "Extracting MNIST_data/t10k-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data/t10k-labels-idx1-ubyte.gz\n",
      "WARNING:tensorflow:From /usr/local/Caskroom/miniconda/base/envs/tensorflow-v1-test/lib/python3.6/site-packages/tensorflow_core/contrib/learn/python/learn/datasets/mnist.py:290: DataSet.__init__ (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use alternatives such as official/mnist/dataset.py from tensorflow/models.\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "mnist = input_data.read_data_sets(\"MNIST_data/\", one_hot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Explore Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_image(image):\n",
    "    plt.imshow(image.reshape((28, 28)), cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label: [0 0 0 1 0 0 0 0 0 0]\n",
      "Argmax: 3\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "index = random.randint(0, len(mnist.train.images))\n",
    "\n",
    "image = mnist.train.images[index]\n",
    "label = mnist.train.labels[index]\n",
    "\n",
    "print('Label: {}\\nArgmax: {}'.format(label.astype(int), np.argmax(label)))\n",
    "\n",
    "show_image(image)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "4829090c-07c5-4964-a84a-701a5b89f294"
    }
   },
   "source": [
    "## Define Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reset graph\n",
    "tf.reset_default_graph()\n",
    "# Set reuse to AUTO_REUSE, this allows you to run tf.get_variable() more than once.\n",
    "tf.get_variable_scope()._reuse = tf.AUTO_REUSE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Learning rate\n",
    "learning_rate = 0.1\n",
    "# Batch Size\n",
    "batch_size = 256\n",
    "# Epochs\n",
    "epochs = 500"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "91eb9701-a56a-4695-ab94-8d9eb2c98206"
    }
   },
   "source": [
    "#### Layer 0: Model Inputs: (28*28)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "nbpresent": {
     "id": "9229f34c-08e4-4285-b2f0-750cdaa195b0"
    }
   },
   "outputs": [],
   "source": [
    "# Input Images\n",
    "features = tf.placeholder(tf.float32, [None, 28*28], name='features')\n",
    "\n",
    "# True Values\n",
    "labels = tf.placeholder(tf.float32, [None, 10], name='labels')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "a12af5e5-d37f-40b3-83e3-4bca21ab0eca"
    }
   },
   "source": [
    "#### Layer 1: Fully Connected: 32"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "nbpresent": {
     "id": "6688439c-cd74-4d5f-9f54-ac8eeb169518"
    }
   },
   "outputs": [],
   "source": [
    "# Weights\n",
    "W_1 = tf.get_variable('W_1', shape=[28*28, 32], initializer=tf.truncated_normal_initializer)\n",
    "# Bias\n",
    "b_1 = tf.get_variable('b_1', shape=[32], initializer=tf.zeros_initializer)\n",
    "\n",
    "# Layer Output\n",
    "layer_1 = tf.matmul(features, W_1) + b_1\n",
    "# Activation Function\n",
    "layer_1 = tf.nn.sigmoid(layer_1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Layer 2: Fully Connected: 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Weights\n",
    "W_2 = tf.get_variable('W_2', shape=[32, 10], initializer=tf.truncated_normal_initializer)\n",
    "# Bias\n",
    "b_2 = tf.get_variable('b_2', shape=[10], initializer=tf.zeros_initializer)\n",
    "\n",
    "# Layer Output\n",
    "layer_2 = tf.matmul(layer_1, W_2) + b_2\n",
    "\n",
    "# Model Predictions\n",
    "logits = layer_2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Loss Function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "nbpresent": {
     "id": "77576169-b0c3-4b16-87fb-1d132a2d715e"
    }
   },
   "outputs": [],
   "source": [
    "# Cross entropy\n",
    "cross_entropys = tf.nn.softmax_cross_entropy_with_logits_v2(labels=labels, logits=logits)\n",
    "\n",
    "# Take average accross batch\n",
    "average_cross_entropy = tf.reduce_mean(cross_entropys)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Optimizer\n",
    "optimizer = tf.train.AdamOptimizer(learning_rate)\n",
    "\n",
    "# Training Step\n",
    "train_step = optimizer.minimize(average_cross_entropy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "correct_prediction = tf.equal(tf.argmax(labels, 1), tf.argmax(logits, 1))\n",
    "\n",
    "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "96345023-61d9-4ed4-bd92-0b2558ae6b6d"
    }
   },
   "source": [
    "## Train Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Saver"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "saver = tf.train.Saver()\n",
    "save_path = './checkpoints/simple.ckpt'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Run Tensorflow Session"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:   0 Train Accuracy: 0.156\n",
      "Epoch: 100 Train Accuracy: 0.918\n",
      "Epoch: 200 Train Accuracy: 0.945\n",
      "Epoch: 300 Train Accuracy: 0.926\n",
      "Epoch: 400 Train Accuracy: 0.930\n",
      "Epoch: 500 Train Accuracy: 0.926\n",
      "Test Accuracy: 0.918\n"
     ]
    }
   ],
   "source": [
    "print_every = 100\n",
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    \n",
    "    for e in range(epochs+1):\n",
    "        \n",
    "        batch_features, batch_labels = mnist.train.next_batch(batch_size)\n",
    "\n",
    "        _, acc = sess.run([train_step, accuracy], feed_dict={\n",
    "            features: batch_features,\n",
    "            labels: batch_labels\n",
    "        })\n",
    "        \n",
    "        if e % print_every == 0:\n",
    "            print('Epoch:{:4} Train Accuracy: {:.3f}'.format(e, acc))\n",
    "    \n",
    "    \n",
    "    acc = sess.run(accuracy, feed_dict={\n",
    "        features: mnist.test.images,\n",
    "        labels: mnist.test.labels\n",
    "    })\n",
    "    print('Test Accuracy: {:.3f}'.format(acc))\n",
    "    \n",
    "    saver.save(sess, save_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "063c912e-a9b6-4142-865b-6eb4c246a65e"
    }
   },
   "source": [
    "## Prediction Example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "nbpresent": {
     "id": "edca66a2-2ab3-4b41-9d7b-12f43d8e6c6a"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from ./checkpoints/simple.ckpt\n",
      "\n",
      "Label: [0 0 0 1 0 0 0 0 0 0]\n",
      "Argmax: 3\n",
      "\n",
      "Logit: [-4  0 -1  9 -7  1 -7 -1  0 -4]\n",
      "Argmax: 3\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, save_path)\n",
    "    \n",
    "    index = random.randint(0, len(mnist.train.images))\n",
    "    image = mnist.train.images[index]\n",
    "    label = mnist.train.labels[index]\n",
    "    \n",
    "    logit = sess.run(logits, feed_dict={\n",
    "        features: [image],\n",
    "        labels: [label]\n",
    "    })[0]\n",
    "    \n",
    "    # This is just to make the print statement nicer\n",
    "    label = label.astype(int)\n",
    "    logit = logit.astype(int)\n",
    "\n",
    "    print('\\nLabel: {}\\nArgmax: {}\\n\\nLogit: {}\\nArgmax: {}'.format(label, np.argmax(label), logit, np.argmax(logit)))\n",
    "    \n",
    "    show_image(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:tensorflow-v1-test]",
   "language": "python",
   "name": "conda-env-tensorflow-v1-test-py"
  },
  "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
}