svhn-keras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Convolution2D, Dropout, Flatten, Dense, Reshape, BatchNormalization, Activation, MaxPooling2D\n",
    "import numpy as np\n",
    "import json\n",
    "from scipy.ndimage import imread\n",
    "from scipy.misc import imresize\n",
    "from skimage import color\n",
    "import random\n",
    "import os\n",
    "from nbutil import imshow_multi\n",
    "from keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeQAAAD8CAYAAACmXNe7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvSc9bgeXAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f86d94b1250>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 5  7 10 10 10]\n",
      "(32, 64)\n"
     ]
    }
   ],
   "source": [
    "NUM_OUTPUT_DIGITS = 5\n",
    "\n",
    "class Dataset(object):\n",
    "    def __init__(self, name):\n",
    "        dir = '../data/svhn/' + name\n",
    "        \n",
    "        filename_labels = json.load(open(os.path.join(dir, 'labels.json')))\n",
    "        filenames = filename_labels.keys()\n",
    "        max_len = NUM_OUTPUT_DIGITS\n",
    "        label_matrix = np.zeros((len(filenames), max_len), dtype=int)\n",
    "        label_matrix.fill(10)\n",
    "        for i, filename in enumerate(filenames):\n",
    "            digits = '0123456789'\n",
    "            for j, digit in enumerate(filename_labels[filename]):\n",
    "                if j < NUM_OUTPUT_DIGITS:\n",
    "                    label_matrix[i, j] = digits.index(digit)\n",
    "        \n",
    "        self.label_matrix = label_matrix\n",
    "        self.filenames = filenames\n",
    "        self.dir = dir\n",
    "    \n",
    "    def rand_pair(self):\n",
    "        i = random.randint(0, len(self.filenames)-1)\n",
    "        filename = os.path.join(self.dir, self.filenames[i])\n",
    "        img = imread(filename)\n",
    "        img = color.rgb2gray(img)\n",
    "        img = imresize(img, [32, 64])\n",
    "        return img / 255.0, self.label_matrix[i]\n",
    "    \n",
    "    def rand_batch(self, size=16):\n",
    "        images = []\n",
    "        labels = []\n",
    "        for _ in xrange(size):\n",
    "            img, label = self.rand_pair()\n",
    "            images.append(img)\n",
    "            # labels.append(self.flat_one_hot(label))\n",
    "            labels.append(label)\n",
    "        return np.array(images), np.array(labels)\n",
    "    \n",
    "    def generator(self, batch_size=8):\n",
    "        def gen():\n",
    "            while True:\n",
    "                imgs, labels = self.rand_batch()\n",
    "                yield np.expand_dims(imgs, -1), np.expand_dims(labels, -1)\n",
    "        return gen()\n",
    "    \n",
    "    def flat_one_hot(self, label):\n",
    "        x = np.zeros((NUM_OUTPUT_DIGITS, 11), dtype=int)\n",
    "        for place, val in enumerate(label):\n",
    "            x[place][val] = 1\n",
    "        return x.reshape(NUM_OUTPUT_DIGITS * 11)\n",
    "\n",
    "d = Dataset('train')\n",
    "# print d.rand_pair()\n",
    "img, label = d.rand_pair()\n",
    "imshow_multi([img])\n",
    "print label\n",
    "print img.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "convolution2d_11 (Convolution2D) (None, 32, 64, 16)    416         convolution2d_input_2[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_6 (BatchNorma (None, 32, 64, 16)    64          convolution2d_11[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "maxpooling2d_6 (MaxPooling2D)    (None, 16, 32, 16)    0           batchnormalization_6[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "dropout_6 (Dropout)              (None, 16, 32, 16)    0           maxpooling2d_6[0][0]             \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_12 (Convolution2D) (None, 16, 32, 32)    12832       dropout_6[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_7 (BatchNorma (None, 16, 32, 32)    128         convolution2d_12[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "maxpooling2d_7 (MaxPooling2D)    (None, 8, 16, 32)     0           batchnormalization_7[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "dropout_7 (Dropout)              (None, 8, 16, 32)     0           maxpooling2d_7[0][0]             \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_13 (Convolution2D) (None, 8, 16, 64)     51264       dropout_7[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_8 (BatchNorma (None, 8, 16, 64)     256         convolution2d_13[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "maxpooling2d_8 (MaxPooling2D)    (None, 4, 8, 64)      0           batchnormalization_8[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "dropout_8 (Dropout)              (None, 4, 8, 64)      0           maxpooling2d_8[0][0]             \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_14 (Convolution2D) (None, 4, 8, 128)     204928      dropout_8[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_9 (BatchNorma (None, 4, 8, 128)     512         convolution2d_14[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "maxpooling2d_9 (MaxPooling2D)    (None, 2, 4, 128)     0           batchnormalization_9[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "dropout_9 (Dropout)              (None, 2, 4, 128)     0           maxpooling2d_9[0][0]             \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_15 (Convolution2D) (None, 2, 4, 256)     819456      dropout_9[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_10 (BatchNorm (None, 2, 4, 256)     1024        convolution2d_15[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "maxpooling2d_10 (MaxPooling2D)   (None, 1, 2, 256)     0           batchnormalization_10[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "dropout_10 (Dropout)             (None, 1, 2, 256)     0           maxpooling2d_10[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "flatten_2 (Flatten)              (None, 512)           0           dropout_10[0][0]                 \n",
      "____________________________________________________________________________________________________\n",
      "dense_4 (Dense)                  (None, 512)           262656      flatten_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "dense_5 (Dense)                  (None, 512)           262656      dense_4[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "dense_6 (Dense)                  (None, 55)            28215       dense_5[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "reshape_2 (Reshape)              (None, 5, 11)         0           dense_6[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "activation_2 (Activation)        (None, 5, 11)         0           reshape_2[0][0]                  \n",
      "====================================================================================================\n",
      "Total params: 1,644,407\n",
      "Trainable params: 1,643,415\n",
      "Non-trainable params: 992\n",
      "____________________________________________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "\n",
    "\"\"\"\n",
    "32 x 64 x 3\n",
    "16 x 32 x 32\n",
    "8 x 16 x 64\n",
    "4 x 8 x 128\n",
    "2 x 4 x 256\n",
    "1 x 2 x 512\n",
    "1024\n",
    "512\n",
    "70\n",
    "10 x 7\n",
    "\"\"\"\n",
    "\n",
    "for i in xrange(5):\n",
    "    channels = 16 * (2 ** i)\n",
    "    kwargs = {}\n",
    "    if i == 0:\n",
    "        kwargs['input_shape'] = [32, 64, 1]\n",
    "    model.add(Convolution2D(channels, 5, 5, activation='relu', border_mode='same', **kwargs))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "    model.add(Dropout(0.7))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(512, activation='relu'))\n",
    "model.add(Dense(512, activation='relu'))\n",
    "model.add(Dense(NUM_OUTPUT_DIGITS * 11))\n",
    "model.add(Reshape((NUM_OUTPUT_DIGITS, 11)))\n",
    "model.add(Activation('softmax'))\n",
    "# model.add(Reshape([NUM_OUTPUT_DIGITS * 11]))\n",
    "\n",
    "model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "print model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "5s - loss: 1.1899 - acc: 0.6203\n",
      "Epoch 2/100\n",
      "5s - loss: 1.1806 - acc: 0.6226\n",
      "Epoch 3/100\n",
      "5s - loss: 1.1878 - acc: 0.6220\n",
      "Epoch 4/100\n",
      "5s - loss: 1.1759 - acc: 0.6271\n",
      "Epoch 5/100\n",
      "5s - loss: 1.1815 - acc: 0.6235\n",
      "Epoch 6/100\n",
      "5s - loss: 1.1848 - acc: 0.6176\n",
      "Epoch 7/100\n",
      "5s - loss: 1.1817 - acc: 0.6254\n",
      "Epoch 8/100\n",
      "5s - loss: 1.1714 - acc: 0.6240\n",
      "Epoch 9/100\n",
      "5s - loss: 1.1817 - acc: 0.6201\n",
      "Epoch 10/100\n",
      "5s - loss: 1.1785 - acc: 0.6225\n",
      "Epoch 11/100\n",
      "5s - loss: 1.1795 - acc: 0.6245\n",
      "Epoch 12/100\n",
      "5s - loss: 1.1699 - acc: 0.6277\n",
      "Epoch 13/100\n",
      "5s - loss: 1.1916 - acc: 0.6230\n",
      "Epoch 14/100\n",
      "5s - loss: 1.1560 - acc: 0.6308\n",
      "Epoch 15/100\n",
      "5s - loss: 1.1677 - acc: 0.6277\n",
      "Epoch 16/100\n",
      "5s - loss: 1.1663 - acc: 0.6290\n",
      "Epoch 17/100\n",
      "5s - loss: 1.1768 - acc: 0.6228\n",
      "Epoch 18/100\n",
      "5s - loss: 1.1693 - acc: 0.6280\n",
      "Epoch 19/100\n",
      "5s - loss: 1.1659 - acc: 0.6285\n",
      "Epoch 20/100\n",
      "5s - loss: 1.1661 - acc: 0.6269\n",
      "Epoch 21/100\n",
      "5s - loss: 1.1653 - acc: 0.6309\n",
      "Epoch 22/100\n",
      "5s - loss: 1.1645 - acc: 0.6298\n",
      "Epoch 23/100\n",
      "5s - loss: 1.1614 - acc: 0.6283\n",
      "Epoch 24/100\n",
      "5s - loss: 1.1622 - acc: 0.6324\n",
      "Epoch 25/100\n",
      "5s - loss: 1.1637 - acc: 0.6288\n",
      "Epoch 26/100\n",
      "5s - loss: 1.1731 - acc: 0.6286\n",
      "Epoch 27/100\n",
      "5s - loss: 1.1641 - acc: 0.6290\n",
      "Epoch 28/100\n",
      "5s - loss: 1.1610 - acc: 0.6323\n",
      "Epoch 29/100\n",
      "5s - loss: 1.1490 - acc: 0.6352\n",
      "Epoch 30/100\n",
      "5s - loss: 1.1587 - acc: 0.6311\n",
      "Epoch 31/100\n",
      "5s - loss: 1.1550 - acc: 0.6294\n",
      "Epoch 32/100\n",
      "5s - loss: 1.1573 - acc: 0.6298\n",
      "Epoch 33/100\n",
      "5s - loss: 1.1641 - acc: 0.6273\n",
      "Epoch 34/100\n",
      "5s - loss: 1.1516 - acc: 0.6357\n",
      "Epoch 35/100\n",
      "5s - loss: 1.1660 - acc: 0.6307\n",
      "Epoch 36/100\n",
      "5s - loss: 1.1488 - acc: 0.6360\n",
      "Epoch 37/100\n",
      "5s - loss: 1.1415 - acc: 0.6357\n",
      "Epoch 38/100\n",
      "5s - loss: 1.1384 - acc: 0.6345\n",
      "Epoch 39/100\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-33-69374871509c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit_generator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0md\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgenerator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m4096\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/keras/models.pyc\u001b[0m in \u001b[0;36mfit_generator\u001b[0;34m(self, generator, samples_per_epoch, nb_epoch, verbose, callbacks, validation_data, nb_val_samples, class_weight, max_q_size, nb_worker, pickle_safe, initial_epoch, **kwargs)\u001b[0m\n\u001b[1;32m    933\u001b[0m                                         \u001b[0mnb_worker\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mnb_worker\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    934\u001b[0m                                         \u001b[0mpickle_safe\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpickle_safe\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 935\u001b[0;31m                                         initial_epoch=initial_epoch)\n\u001b[0m\u001b[1;32m    936\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    937\u001b[0m     def evaluate_generator(self, generator, val_samples,\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/keras/engine/training.pyc\u001b[0m in \u001b[0;36mfit_generator\u001b[0;34m(self, generator, samples_per_epoch, nb_epoch, verbose, callbacks, validation_data, nb_val_samples, class_weight, max_q_size, nb_worker, pickle_safe, initial_epoch)\u001b[0m\n\u001b[1;32m   1555\u001b[0m                     outs = self.train_on_batch(x, y,\n\u001b[1;32m   1556\u001b[0m                                                \u001b[0msample_weight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msample_weight\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1557\u001b[0;31m                                                class_weight=class_weight)\n\u001b[0m\u001b[1;32m   1558\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1559\u001b[0m                     \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/keras/engine/training.pyc\u001b[0m in \u001b[0;36mtrain_on_batch\u001b[0;34m(self, x, y, sample_weight, class_weight)\u001b[0m\n\u001b[1;32m   1318\u001b[0m             \u001b[0mins\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mx\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0my\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0msample_weights\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1319\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_make_train_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1320\u001b[0;31m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1321\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1322\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/keras/backend/tensorflow_backend.pyc\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   1941\u001b[0m         \u001b[0msession\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_session\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1942\u001b[0m         updated = session.run(self.outputs + [self.updates_op],\n\u001b[0;32m-> 1943\u001b[0;31m                               feed_dict=feed_dict)\n\u001b[0m\u001b[1;32m   1944\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mupdated\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1945\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    765\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    766\u001b[0m       result = self._run(None, fetches, feed_dict, options_ptr,\n\u001b[0;32m--> 767\u001b[0;31m                          run_metadata_ptr)\n\u001b[0m\u001b[1;32m    768\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    769\u001b[0m         \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_run\u001b[0;34m(self, handle, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    963\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfinal_fetches\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mfinal_targets\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    964\u001b[0m       results = self._do_run(handle, final_targets, final_fetches,\n\u001b[0;32m--> 965\u001b[0;31m                              feed_dict_string, options, run_metadata)\n\u001b[0m\u001b[1;32m    966\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    967\u001b[0m       \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_do_run\u001b[0;34m(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m   1013\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mhandle\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1014\u001b[0m       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,\n\u001b[0;32m-> 1015\u001b[0;31m                            target_list, options, run_metadata)\n\u001b[0m\u001b[1;32m   1016\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1017\u001b[0m       return self._do_call(_prun_fn, self._session, handle, feed_dict,\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_do_call\u001b[0;34m(self, fn, *args)\u001b[0m\n\u001b[1;32m   1020\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m_do_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1021\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1022\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1023\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0merrors\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mOpError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1024\u001b[0m       \u001b[0mmessage\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcompat\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_text\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc\u001b[0m in \u001b[0;36m_run_fn\u001b[0;34m(session, feed_dict, fetch_list, target_list, options, run_metadata)\u001b[0m\n\u001b[1;32m   1002\u001b[0m         return tf_session.TF_Run(session, options,\n\u001b[1;32m   1003\u001b[0m                                  \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_list\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1004\u001b[0;31m                                  status, run_metadata)\n\u001b[0m\u001b[1;32m   1005\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1006\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_prun_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msession\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhandle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "model.fit_generator(d.generator(), 4096, 100, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeQAAAD8CAYAAACmXNe7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f865b1662d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 2  1 10 10 10]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "img, label = d.rand_pair()\n",
    "img_batch = np.expand_dims(np.expand_dims(img, 0), -1)\n",
    "pred = model.predict(img_batch).argmax(axis=-1)\n",
    "imshow_multi([img])\n",
    "print pred[0]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12+"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}