FMAI_Chapter9-1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Successfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.\n",
      "Extracting ./data/mnist/train-images-idx3-ubyte.gz\n",
      "Successfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.\n",
      "Extracting ./data/mnist/train-labels-idx1-ubyte.gz\n",
      "Successfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.\n",
      "Extracting ./data/mnist/t10k-images-idx3-ubyte.gz\n",
      "Successfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.\n",
      "Extracting ./data/mnist/t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "\n",
    "mnist = input_data.read_data_sets(\"./data/mnist/\", one_hot = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "num_input = 28 * 28\n",
    "num_hidden = 256\n",
    "\n",
    "learning_rate = 0.01\n",
    "training_epoch = 10\n",
    "batch_size = 100\n",
    "\n",
    "num_output_sample = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x_true = tf.placeholder(tf.float32, [None, num_input])\n",
    "y_true = x_true\n",
    "\n",
    "weight_encoder = tf.Variable(tf.truncated_normal([num_input, num_hidden]))\n",
    "bias_encoder = tf.Variable(tf.truncated_normal([num_hidden]))\n",
    "\n",
    "weight_decoder = tf.Variable(tf.truncated_normal([num_hidden, num_input]))\n",
    "bias_decoder = tf.Variable(tf.truncated_normal([num_input]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "encoder = tf.nn.sigmoid(tf.add(tf.matmul(x_true, weight_encoder), bias_encoder))\n",
    "decoder = tf.nn.sigmoid(tf.add(tf.matmul(encoder, weight_decoder), bias_decoder))\n",
    "\n",
    "cost = tf.reduce_mean(tf.square(y_true - decoder))\n",
    "optimizer = tf.train.AdamOptimizer(learning_rate)\n",
    "optimizer = optimizer.minimize(cost)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sess = tf.Session()\n",
    "sess.run(tf.global_variables_initializer())\n",
    "\n",
    "total_batch = mnist.train.num_examples // batch_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch : 1, Cost : 0.0601339857483452\n",
      "Epoch : 2, Cost : 0.03704362856393511\n",
      "Epoch : 3, Cost : 0.031147710738534277\n",
      "Epoch : 4, Cost : 0.02941297791220925\n",
      "Epoch : 5, Cost : 0.02824928531931205\n",
      "Epoch : 6, Cost : 0.027675347724421456\n",
      "Epoch : 7, Cost : 0.02728964924812317\n",
      "Epoch : 8, Cost : 0.02705718839371746\n",
      "Epoch : 9, Cost : 0.026880800398913298\n",
      "Epoch : 10, Cost : 0.026718920770693908\n",
      "Done\n"
     ]
    }
   ],
   "source": [
    "for epoch in range(training_epoch):\n",
    "    total_cost = 0\n",
    "    for i in range(total_batch):\n",
    "        batch_xs, batch_ys = mnist.train.next_batch(batch_size)\n",
    "        _, cost_value = sess.run([optimizer, cost], {x_true : batch_xs})\n",
    "        total_cost += cost_value\n",
    "    print(\"Epoch : {0}, Cost : {1}\".format(epoch + 1, total_cost / total_batch))\n",
    "print(\"Done\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAACQCAYAAADKiyWfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0rHHjuFlN3gAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x22d16089c88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlMAAACNCAYAAACT6v+eAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x22d17d184e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "samples = sess.run(decoder, {x_true : mnist.test.images[:num_output_sample]})\n",
    "fig, ax = plt.subplots(2, num_output_sample, figsize = (num_output_sample, 2))\n",
    "\n",
    "for i in range(num_output_sample):\n",
    "    ax[0][i].set_axis_off\n",
    "    ax[1][i].set_axis_off\n",
    "    ax[0][i].imshow(np.reshape(mnist.test.images[i], (28, 28)))\n",
    "    ax[1][i].imshow(np.reshape(samples[i], (28, 28)))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}