Tensorflow - MNIST-Fashion.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting mnist-fashion\\train-images-idx3-ubyte.gz\n",
      "Extracting mnist-fashion\\train-labels-idx1-ubyte.gz\n",
      "Extracting mnist-fashion\\t10k-images-idx3-ubyte.gz\n",
      "Extracting mnist-fashion\\t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "#Import\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import random\n",
    "from matplotlib import pyplot as plt\n",
    "import matplotlib.image as mpimg\n",
    "#np.set_printoptions(threshold=np.nan)\n",
    "\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "data = input_data.read_data_sets('mnist-fashion')\n",
    "\n",
    "#Dataset\n",
    "#mnist = tf.contrib.learn.datasets.load_dataset(\"mnist\")\n",
    "train_data = data.train.images # Returns np.array\n",
    "ou_train_labels = np.asarray(data.train.labels, dtype=np.int32)\n",
    "test_data = data.test.images # Returns np.array\n",
    "ou_test_labels = np.asarray(data.test.labels, dtype=np.int32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fmnist_name = {\n",
    "    0: \"T-Shirt/Top\",\n",
    "    1: \"Trouser\",\n",
    "    2: \"Pullover\",\n",
    "    3: \"Dress\",\n",
    "    4: \"Coat\",\n",
    "    5: \"Sandal\",\n",
    "    6: \"Shirt\",\n",
    "    7: \"Sneaker\",\n",
    "    8: \"Bag\",\n",
    "    9: \"Ankle boot\"\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def weight_variable(shape):\n",
    "    return tf.Variable(tf.truncated_normal(shape, stddev=0.1))\n",
    "def bias_variable(shape):\n",
    "    return tf.Variable(tf.constant(0.1, shape=shape))\n",
    "\n",
    "def conv2d(x, W):\n",
    "    return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')\n",
    "def max_pool_2x2(x):\n",
    "    return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],\n",
    "                        strides=[1, 2, 2, 1], padding='SAME')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def next_batch(num, data, labels):\n",
    "    '''\n",
    "    Return a total of `num` random samples and labels. \n",
    "    '''\n",
    "    idx = np.arange(0 , len(data))\n",
    "    np.random.shuffle(idx)\n",
    "    idx = idx[:num]\n",
    "    data_shuffle = [data[ i] for i in idx]\n",
    "    labels_shuffle = [labels[ i] for i in idx]\n",
    "\n",
    "    return np.asarray(data_shuffle), np.asarray(labels_shuffle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_labels = []\n",
    "test_labels = []\n",
    "\n",
    "for i in range(len(ou_train_labels)):\n",
    "    label_dat = np.zeros(10)\n",
    "    label_dat[ou_train_labels[i]] = 1\n",
    "\n",
    "    train_labels.append(label_dat)\n",
    "    \n",
    "for i in range(len(ou_test_labels)):\n",
    "    label_dat = np.zeros(10)\n",
    "    label_dat[ou_test_labels[i]] = 1\n",
    "\n",
    "    test_labels.append(label_dat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x = tf.placeholder(tf.float32, [None, 784])\n",
    "\n",
    "x_image = tf.reshape(x, [-1, 28, 28, 1])\n",
    "W_conv1 = weight_variable([5, 5, 1, 32])\n",
    "b_conv1 = bias_variable([32])\n",
    "\n",
    "h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)\n",
    "h_pool1 = max_pool_2x2(h_conv1)\n",
    "\n",
    "W_conv2 = weight_variable([5, 5, 32, 64])\n",
    "b_conv2 = bias_variable([64])\n",
    "\n",
    "h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)\n",
    "h_pool2 = max_pool_2x2(h_conv2)\n",
    "\n",
    "W_fc1 = weight_variable([7 * 7 * 64, 1024])\n",
    "b_fc1 = bias_variable([1024])\n",
    "\n",
    "h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])\n",
    "h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)\n",
    "\n",
    "keep_prob = tf.placeholder(tf.float32)\n",
    "h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)\n",
    "\n",
    "W_fc2 = weight_variable([1024, 64])\n",
    "b_fc2 = bias_variable([64])\n",
    "\n",
    "h_fc2 = tf.matmul(h_fc1_drop, W_fc2) + b_fc2\n",
    "\n",
    "W_fc3 = weight_variable([64, 10])\n",
    "b_fc3 = bias_variable([10])\n",
    "\n",
    "y = tf.matmul(h_fc2, W_fc3) + b_fc3\n",
    "\n",
    "result = tf.nn.softmax(y)\n",
    "\n",
    "y_ = tf.placeholder(tf.float32, [None, 10])\n",
    "\n",
    "cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y))\n",
    "train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "init = tf.global_variables_initializer()\n",
    "\n",
    "sess = tf.InteractiveSession()\n",
    "sess.run(init)\n",
    "\n",
    "saver = tf.train.Saver()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Load the model\n",
    "saver.restore(sess, \"./tf-models/mnist-fashion.ckpt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test error 0.42844, train error 0.6389828\n",
      "test error 0.420159, train error 0.46614"
     ]
    }
   ],
   "source": [
    "try:    \n",
    "    while True:\n",
    "        test_batch =next_batch(25, test_data, test_labels)\n",
    "        train_batch = next_batch(25, train_data, train_labels)\n",
    "        te = cross_entropy.eval(feed_dict={x: test_batch[0], y_: test_batch[1], keep_prob: 1.0})\n",
    "        tee = cross_entropy.eval(feed_dict={x: train_batch[0], y_: train_batch[1], keep_prob: 1.0})\n",
    "        print(\"\\rtest error %g, train error %g\"%(te, tee), end=\"\")\n",
    "        for i in range(11):\n",
    "            batch = next_batch(100, train_data, train_labels)\n",
    "            sess.run(train_step, feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.4})\n",
    "except KeyboardInterrupt:\n",
    "    print(\"\")\n",
    "test_batch =next_batch(75, test_data, test_labels)\n",
    "train_batch = next_batch(75, train_data, train_labels)\n",
    "te = cross_entropy.eval(feed_dict={x: test_batch[0], y_: test_batch[1], keep_prob: 1.0})\n",
    "tee = cross_entropy.eval(feed_dict={x: train_batch[0], y_: train_batch[1], keep_prob: 1.0})\n",
    "print(\"test error %g, train error %g\"%(te, tee), end=\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'./tf-models/mnist-fashion.ckpt'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Save the model\n",
    "saver.save(sess, \"./tf-models/mnist-fashion.ckpt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFMNPFBTDTxQUw08U1P8DnOFavgTV6sQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x4d3c278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "suggested label: Coat\n",
      "confidence: 51%\n",
      "correct label: Coat\n",
      "T-Shirt/Top     0.00\t 0\n",
      "Trouser         0.00\t 0\n",
      "Pullover        0.38\t 0\n",
      "Dress           0.00\t 0\n",
      "Coat            0.51\t 1\n",
      "Sandal          0.00\t 0\n",
      "Shirt           0.11\t 0\n",
      "Sneaker         0.00\t 0\n",
      "Bag             0.00\t 0\n",
      "Ankle boot      0.00\t 0\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvG+/AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xc1eea58>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "suggested label: Sneaker\n",
      "confidence: 100%\n",
      "correct label: Sneaker\n",
      "T-Shirt/Top     0.00\t 0\n",
      "Trouser         0.00\t 0\n",
      "Pullover        0.00\t 0\n",
      "Dress           0.00\t 0\n",
      "Coat            0.00\t 0\n",
      "Sandal          0.00\t 0\n",
      "Shirt           0.00\t 0\n",
      "Sneaker         1.00\t 1\n",
      "Bag             0.00\t 0\n",
      "Ankle boot      0.00\t 0\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nICjCDwRF+IGgCD8QFOEHgiL8QFB/Au/4RVQgwUuKAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xd115d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "suggested label: Sneaker\n",
      "confidence: 97%\n",
      "correct label: Sneaker\n",
      "T-Shirt/Top     0.00\t 0\n",
      "Trouser         0.00\t 0\n",
      "Pullover        0.00\t 0\n",
      "Dress           0.00\t 0\n",
      "Coat            0.00\t 0\n",
      "Sandal          0.01\t 0\n",
      "Shirt           0.00\t 0\n",
      "Sneaker         0.97\t 1\n",
      "Bag             0.00\t 0\n",
      "Ankle boot      0.02\t 0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "image = random.randint(0,len(test_data)-5)\n",
    "\n",
    "#num = int(input(\"How many pictues shall be shown?\"))\n",
    "num = 3\n",
    "res = sess.run(result, feed_dict={x: test_data[image:image+num], keep_prob: 1})\n",
    "\n",
    "#print(test_data[0])\n",
    "\n",
    "for j,e in enumerate(res):\n",
    "    data = []\n",
    "    for i in range(28):\n",
    "        tmp = []\n",
    "        for k in range(28):\n",
    "            tmp.append([test_data[image+j][28*i+k-1],test_data[image+j][28*i+k-1],test_data[image+j][28*i+k-1]])\n",
    "        data.append(tmp)\n",
    "    dat = np.array(data, dtype=np.float32)\n",
    "    plt.imshow(dat)\n",
    "    plt.show()\n",
    "    \n",
    "    print(\"suggested label: {:s}\\nconfidence: {:.0f}%\\ncorrect label: {:s}\".format(fmnist_name[e.argmax()], e.max()*100, fmnist_name[test_labels[image+j].argmax()]))\n",
    "    \n",
    "    for i in range(10):\n",
    "        print(\"{:15s} {:1.2f}\\t {:1.0f}\".format(fmnist_name[i], e[i], test_labels[image+j][i]))\n",
    "    print()"
   ]
  },
  {
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}