00-MNIST-Data-Basic-Approach-SA-Udemy-TF.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# MNIST Data Set - Basic Approach"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Get the MNIST Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.examples.tutorials.mnist import input_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-7-758d29429358>:1: 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 C:\\Users\\saga54\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\tensorflow\\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 C:\\Users\\saga54\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\tensorflow\\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 C:\\Users\\saga54\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\tensorflow\\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 C:\\Users\\saga54\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\tensorflow\\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 C:\\Users\\saga54\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\tensorflow\\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": [
    "mnist = input_data.read_data_sets(\"MNIST_data/\",one_hot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "** Alternative sources of the data just in case: **\n",
    "\n",
    "* http://yann.lecun.com/exdb/mnist/\n",
    "* https://github.com/mrgloom/MNIST-dataset-in-different-formats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensorflow.contrib.learn.python.learn.datasets.base.Datasets"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(mnist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tensorflow.contrib.learn.python.learn.datasets.mnist.DataSet at 0xf71fb70>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(55000, 784)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train.images.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "55000"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train.num_examples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10000"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.test.num_examples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5000"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.validation.num_examples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualizing the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(784,)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train.images[1].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.3137255 , 0.46274513,\n",
       "        0.46274513, 0.46274513, 0.64705884, 1.        , 0.8941177 ,\n",
       "        0.46274513, 0.46274513, 0.46274513, 0.02352941, 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.14509805, 0.8313726 , 0.9176471 , 0.9686275 , 0.9921569 ,\n",
       "        0.9921569 , 0.9921569 , 0.9921569 , 0.9921569 , 0.9921569 ,\n",
       "        0.9921569 , 0.9921569 , 0.9921569 , 0.9215687 , 0.9176471 ,\n",
       "        0.49411768, 0.1137255 , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.3803922 , 0.9921569 , 0.9921569 , 0.9921569 , 0.9921569 ,\n",
       "        0.9921569 , 0.86274517, 0.69411767, 0.69411767, 0.69411767,\n",
       "        0.69411767, 0.86666673, 0.9921569 , 0.9921569 , 0.9921569 ,\n",
       "        0.9921569 , 0.882353  , 0.31764707, 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.09411766, 0.6901961 , 0.77647066, 0.77647066, 0.77647066,\n",
       "        0.77647066, 0.14901961, 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.13725491, 0.2392157 , 0.45882356, 0.8470589 ,\n",
       "        0.9921569 , 0.9921569 , 0.37647063, 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.05882353, 0.7058824 ,\n",
       "        0.9921569 , 0.9921569 , 0.37647063, 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.7254902 , 0.9921569 ,\n",
       "        0.9921569 , 0.86666673, 0.15294118, 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.25490198, 0.9686275 , 0.9921569 ,\n",
       "        0.9921569 , 0.29411766, 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.5529412 , 0.9803922 , 0.9921569 , 0.9921569 ,\n",
       "        0.52156866, 0.00784314, 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.2901961 , 0.9490197 , 0.9921569 , 0.9921569 , 0.50980395,\n",
       "        0.02352941, 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.15686275,\n",
       "        0.9294118 , 0.9921569 , 0.9921569 , 0.22352943, 0.02745098,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.58431375,\n",
       "        0.9921569 , 0.9921569 , 0.63529414, 0.01568628, 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.5254902 , 0.90196085,\n",
       "        0.9921569 , 0.9176471 , 0.18039216, 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.21568629, 0.91372555, 0.9921569 ,\n",
       "        0.98823535, 0.4156863 , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.21176472, 0.94117653, 0.9921569 , 0.90196085,\n",
       "        0.454902  , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.43529415, 0.9921569 , 0.9921569 , 0.80392164,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.43529415, 0.9921569 , 0.9921569 , 0.3372549 ,\n",
       "        0.        , 0.18823531, 0.5882353 , 0.86274517, 0.2901961 ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.12941177, 0.86274517, 0.9921569 , 0.64705884, 0.23529413,\n",
       "        0.5921569 , 0.90196085, 0.9921569 , 0.7607844 , 0.5058824 ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.03921569,\n",
       "        0.85098046, 0.9921569 , 0.9921569 , 0.75294125, 0.91372555,\n",
       "        0.9921569 , 0.8862746 , 0.29411766, 0.01568628, 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.0509804 ,\n",
       "        0.95294124, 0.9921569 , 0.9921569 , 0.9921569 , 0.97647065,\n",
       "        0.909804  , 0.39607847, 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.23529413, 0.6862745 , 0.9921569 , 0.8470589 , 0.36862746,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ],\n",
       "       [0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        , 0.        , 0.        ,\n",
       "        0.        , 0.        , 0.        ]], dtype=float32)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train.images[7000].reshape(28,28)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x117e03c8>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11761518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(mnist.train.images[7000].reshape(28,28))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x11838f28>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117f8a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(mnist.train.images[1000].reshape(28,28),cmap='gist_gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train.images[1].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.train.images[1].min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x21083579908>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAADQAAAD8CAYAAAA4w4cyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x21083542278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(mnist.train.images[1].reshape(784,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x210835c0128>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAD4AAAD8CAYAAAAv4Rf7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAACk5JREFUeJztnV9sHMUdxz/fOATXhEISE2NK3Bg1kNiVTHOBFkolJEoJ\naUV4QFWAB6RGQkGlAvWhxEFU6gMVTSVIHkqliNIGRYBMKyiiLRBQqzQSOE7SBDDEkEAQWAEnIKAE\n9Y/dXx9mLixO7Nvd8/n2mPlIq9ub3bmZ7+3O7O58Z2ZlZoTIjHpnoF5E4aERhYdGFD7VSFouaUjS\nfklra5VOXlSL67ikJuBV4HLgbWAAuNbMXp7yxHJSqyN+IbDfzF43s/8ADwMra5RWLmol/EvAW4nv\nb/uwY0i6UdJOv1jO5XDeDM7MG7FazGwTsAlAUt7y9mbe9Gt1xIeBBYnvZ/uwwlAr4QPAIkmdkmYB\nq4DHa5RWLmpyqpvZqKSbgaeAJuB+MxusRVp5qcnlLHMm8pfxXWa2LE/EeOcWGlF4aEThoRGFh0YU\nHhpReGhE4aERhYdGFB4aUfhESLpf0oiklxJhcyVtlfSa/5yT2Nbr/bIhSVfUKuPVkuaI/w5YPi5s\nLfCsmS0CnvXfkdSFa0ru9nHu9T5a4ago3My2Ae+PC14JbPbrm4GrE+EPm9m/zewNYD/ORyscect4\nm5kd8uvvAG1+vaJnVibpneXMQ1VUbSiYmeVpF58i7yw3eY/4u5LaAfzniA8vvGdWJq/wx4Eb/PoN\nwB8T4asknSypE1gE7KguizXCzCZdgIeAQ8B/cWV2NTAPV5u/BjwDzE3sfztwABgCrqz0+z6O5Vx2\npvn9Ey3ROwuNKDw0ovDQiMJDIwoPjUIIL5VKjI2NZV6qId6yhkYUHhpReGjUraN+kp6eHp5++unM\n8dra2irvNAGFED5z5kxaW1unN81pTW0C9u7dW9XRy0WKhsAFwF+Bl4FB4BYfPhfYimtw3ArMScTp\nxbkoQ8AVRWxsTCO8HVjq10/FjSfrAtYDa334WuAXfr0L2AucDHTiWlybiiY8jXd2yMx2+/V/Aq/g\nbKGG9s8yXc4kLQS+BvRTpX/WMN6ZpNnAH4BbzewjSce25fHPkt5Ze3u7rV69Okt0AO68887McZIZ\nSFPOT8KNKPpxImwIaE/UA0OJiq03sd9TwEWT/X6pVLKxsbHMC1WU8YpHXO7Q/gZ4xczuTmwq+2d3\ncbx/9qCku4GzSOGfffLJJ+zevbviQZpSUhztS3A16AvAHr+sYAr9M6J3lpnYEJGVKDw0CiG8Hq2s\nwT6dFUJ4c3Mzixcvzhxv+/btudOMl7PQiMJDI1jhhajVlyxZwgMPPJA53gUXXJA7zUIIP3r0KLt2\n7ZreRPM+1k3lUiqVLA/Ex9LsBFu5ReGhkWb4VbOkHZL2ShqU9DMf3thDsFI0BAqYbZ82M/cD36DB\nLaSK13F/2fjYfz3JL4azii714ZuBvwG3kbCQgDcklS2k5yZKo6WlJddjaTVN0qluYPyguV3AV4Bf\nmVm/pMkspOcT0SccglVmyZIl9Pf3Z8o4OF89L6limtkYcL6k04FHJX113PbMFpKkG4Ebj2WkChF5\nyFSrm9kHOK98OVUOwTKzTWa2LO8NSLWkqdXP8EcaSV/AzbW6jwYfgpXm/GoHNvtyPgPoM7MnJD0H\n9ElajZst8/sAZjYoqQ/Xg2IU+KEvKoUi3quHRiGex+fPn8/111+fOd4999yTO814qodGFB4awQov\nRK0+e/Zsli5dmjnetm3bcqdZCOGjo6OMjIxU3nEKKYTw7u7uaX8sjdfx0IjCQyNY4YWo1Ts6Oli3\nbl3meGvWrMmdZiFq9WXLltnAwEDmeDNmzMhdqxfiiA8PD3PHHXdMa5qFOOKFvo5LapL0D0lP+O8N\n7Z1lqdVvwY1AKtPQ0xemEi7pbOC7wH2J4CCGX20AfgL8LxHW0MOv0jgp3wNGzGzCbkneUc08/Kps\nIZVKJUZHRzMv1ZDmcvZN4CpJK4Bm4IuStuC9MzM71JDTF2Yx03F++BN+/Zd8tmPAer/ezWc7BrxO\nI3YMmIS7iN5ZlZko8g3M540oPDQK8XQWbLftlpaWXIZCNcRaPTSi8NCIwkOjEJez5uZmFi5cmDne\nvn37cqdZCOGdnZ3TfgMTr+OhEYWHRhQeGmmdlIOSXpS0p2wANLp3lupyJukgsMzMjiTC1gPvm9ld\nktbi5my8zXtnD+Fso7NwE2CdO1lLa1tbm1133XWZM79hw4bcl7O07ekHgdZxYVM2nxtFnLOx/P8A\nz0ja5YdNQYN7Z2lvWS8xs2FJ84Gtkj5zk5xn3Jk1wmu/zGzYf44Aj+LKb0O/+iuNW3qKpFPL68B3\ngJcIYNxZG25YZXn/B83sSUkDRO+sykzEp7PpIwoPjWCFF6LNrbu7m0ceeSRzvK6urtxpxlo9NApz\nqvf19eWKl5d4qodGFB4ahajcOjo66O3tzRzvpptuyp1mrNxCoxDC48scsxNP9axE4ZMh6XRJv5e0\nT9Irki4KxTvbDPzdzO6TNAtoAdYxRd7ZvHnzbMWKFZkzv2XLltp5Z8BpwBv4Pykk76wTOAz81g+x\nvM8bC1PmnXV0dBTyrRkzgaXAj8xNUrkRP5yyTLXe2YIFC2zjxo1ZoldPilP9TOBg4vu3gD/xeT/V\nzewd4C1J5/mgy3D2UEN7Z2k7BpwP7MS9+usxYA7xtV/VE+wta7DvQhodHeXIkSOVd5xCCiG8Hu8m\njmU8NKLw0ChM5TZnzpzKO47j8OHD+dPMHXMK6enpYceO7He1TU35p5eJp3o9OXDgANdcc820phmv\n46ERhYdGFF5Pgn0eHxwcrKoHUy5StIedB+xJLB8BtwJzga24NretOCelHKcXN4HdEHBFEdvcsu0M\nTTjz4Ms0+Gu/spbxy4ADZvYmgUxdWGYVzhCEKi2kepNletJZwFXAcd2MzZ2vmW476z3uLMsRvxLY\nbWbv+u+f79d+JbiWT09zCMRCOgV4DzgtERYtpGqJj6XTSBQeGlF4aEThoRGFh0YUHhpReGhE4aER\nhYdGsMIL4Z0BH+Pa505EKzBRR9fzJgivSFGED03UdiZp52Tb8iYY7KkehdeZTTXYNimFaFevB0U5\n4tNOFD6dnGgksqTlfvTxfj9INzkD8AFJ/5L0gaQjkl6QtDTxe5dK+tDPFLxH0k8rZiKv6VbNwvHd\nSNbjTMZzgFm4riRduIlw5/ttPwCe9NtWAf2J37sU/wK6tEu9bmBW+syC60byPLDPzF4HkPSw3wfc\nuNb9wNf9vgtxfWtOL79aME8G6lXGx3cjaeXE3UcM+DVQAi7y+5S3je9icrEvAn+RVLHvWM2OuKRn\ncANyx3N78ovZpCORL8EJvhrXDaUH191sPLuBDjP72L908jFch4QJqZlwM/v2RNskjX8D5nucoPuI\nuWmPh4EzgFdxf8IgrmvJ5f4TMzv2Z5jZnyXdK6nVErODj6dep/r4biR9wCJJnb6T0SrcHM+nAgPA\nubiZCxb7bQeBD8vFRdKZ8jPmSroQp+u9SXNQp1r9uG4kwApc7X0UVxzOwZXjt3Ez/R4BPvSCXgR+\nDqzxv3cz7kzYi6soL26IriD1IN65hUYUHhpReGgEK/z/BjqgVdjEuusAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x21083502ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(mnist.train.images[1].reshape(784,1),cmap='gist_gray',aspect=0.02)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = tf.placeholder(tf.float32,shape=[None,784])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 10 because 0-9 possible numbers\n",
    "W = tf.Variable(tf.zeros([784,10]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "b = tf.Variable(tf.zeros([10]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the Graph\n",
    "y = tf.matmul(x,W) + b "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loss and Optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_true = tf.placeholder(tf.float32,[None,10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Cross Entropy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-33-3c5e42f2a280>:1: softmax_cross_entropy_with_logits (from tensorflow.python.ops.nn_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "\n",
      "Future major versions of TensorFlow will allow gradients to flow\n",
      "into the labels input on backprop by default.\n",
      "\n",
      "See @{tf.nn.softmax_cross_entropy_with_logits_v2}.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y_true, logits=y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = optimizer.minimize(cross_entropy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Session"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "init = tf.global_variables_initializer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9217\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    sess.run(init)\n",
    "    \n",
    "    # Train the model for 1000 steps on the training set\n",
    "    # Using built in batch feeder from mnist for convenience\n",
    "    \n",
    "    for step in range(1000):\n",
    "        \n",
    "        batch_x , batch_y = mnist.train.next_batch(100)\n",
    "        \n",
    "        sess.run(train,feed_dict={x:batch_x,y_true:batch_y})\n",
    "        \n",
    "    # Test the Train Model\n",
    "    matches = tf.equal(tf.argmax(y,1),tf.argmax(y_true,1))\n",
    "    \n",
    "    acc = tf.reduce_mean(tf.cast(matches,tf.float32))\n",
    "    \n",
    "    print(sess.run(acc,feed_dict={x:mnist.test.images,y_true:mnist.test.labels}))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While this may seem pretty good, we can actually do much better, the best models can get above 99% accuracy.\n",
    "\n",
    "How do they do this? By using other models, such as convolutional neural networks!"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}