scene-recognition-pretrained.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras import models, layers, applications, preprocess_input\n",
    "from keras.preprocessing import image\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "from scipy.ndimage import imread\n",
    "from scipy.misc import imresize\n",
    "\n",
    "# from keras_models.imagenet_utils import preprocess_input\n",
    "# from keras_models.inception_v3 import InceptionV3\n",
    "\n",
    "from nbutil import imshow_multi\n",
    "from PIL import Image\n",
    "\n",
    "import random\n",
    "import os"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load the data\n",
    "\n",
    "Load existing training and test data, and resize to 64x64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data_dir = '../data/proj6/data/15SceneData'\n",
    "categories = [c for c in os.listdir(os.path.join(data_dir, 'train')) if c[0] != '.']\n",
    "\n",
    "SIZE = 139\n",
    "\n",
    "def load_dataset(name):\n",
    "    images = []\n",
    "    labels = []\n",
    "    for cat in categories:\n",
    "        for filename in os.listdir(os.path.join(data_dir, name, cat)):\n",
    "            if filename.endswith('.jpg'): \n",
    "                path = os.path.join(data_dir, name, cat, filename)\n",
    "                image = imread(path)\n",
    "                h,w = image.shape[:2]\n",
    "                # top-left crop\n",
    "                if h > w:\n",
    "                    image = image[:w, :]\n",
    "                else:\n",
    "                    image = image[:, :h]\n",
    "                image = imresize(image, size=[SIZE, SIZE], mode='L')\n",
    "                images.append(image)\n",
    "                labels.append(categories.index(cat))\n",
    "    images = np.expand_dims(np.stack(images) / 255.0, -1)\n",
    "    labels = np.expand_dims(np.stack(labels).astype(int), -1)\n",
    "    return images, labels\n",
    "\n",
    "trainx, trainy = load_dataset('train')\n",
    "testx, testy = load_dataset('test')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "15"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_dim = max(trainy)[0] + 1\n",
    "y_dim"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Do some jittering and centering\n",
    "\n",
    "Use Keras's `ImageDataGenerator` to apply image jitter. Compute the mean image using the training data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "image_gen = ImageDataGenerator(rotation_range=5, \n",
    "                               zoom_range=0.2, \n",
    "                               width_shift_range=0.1, \n",
    "                               height_shift_range=0.1, \n",
    "                               horizontal_flip=True,\n",
    "                               channel_shift_range=0.2)\n",
    "# samplewise centering is broken in keras (https://github.com/fchollet/keras/issues/4047)\n",
    "# so do it manually:\n",
    "# mean_img = np.mean(trainx, axis=0)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Subtract the mean image and convert all to RGB\n",
    "\n",
    "(since InceptionV3 takes 3-channel inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def inceptionv3_preprocess(x):\n",
    "    # recenter around 0:\n",
    "    x = (x - 0.5) * 2\n",
    "    # inception wants 3 channels, not 1:\n",
    "    x = np.repeat(x, 3, axis=-1)\n",
    "    return x\n",
    "\n",
    "trainx_rgb = inceptionv3_preprocess(trainx - mean_img)\n",
    "testx_rgb = inceptionv3_preprocess(testx - mean_img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1500, 128, 128, 3)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainx_rgb.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load a pre-trained network\n",
    "\n",
    "Pretrained `VGG-F` networks aren't available for Keras, so use `InceptionV3` instead."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "input = layers.Input(shape=(SIZE, SIZE, 3))\n",
    "model = applications.InceptionV3(weights='imagenet', include_top=False, input_tensor=input)\n",
    "# model = InceptionV3(weights='imagenet', include_top=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create a new model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(?, 6, 6, 768)\n",
      "(?, 6, 6, 256)\n",
      "(?, 1, 1, 256)\n",
      "(?, ?)\n"
     ]
    }
   ],
   "source": [
    "output = model.get_layer('mixed5').get_output_at(0)\n",
    "print(output.shape)\n",
    "output = layers.Conv2D(256, (1,1), activation='elu')(output)\n",
    "print(output.shape)\n",
    "output = layers.Dropout(0.5)(output)\n",
    "output = layers.AveragePooling2D((6,6))(output)\n",
    "print(output.shape)\n",
    "output = layers.Flatten()(output)\n",
    "print(output.shape)\n",
    "output = layers.Dense(y_dim, activation='softmax')(output)\n",
    "model2 = models.Model(inputs=[input], outputs=[output])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5, 15)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = model2.predict(trainx_rgb[:5])\n",
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2.compile(loss=keras.losses.sparse_categorical_crossentropy,\n",
    "              optimizer=keras.optimizers.Adam(lr=1e-4),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "188/187 [==============================] - 25s - loss: 0.5199 - acc: 0.8504    \n",
      "epoch: 1, loss: 0.5351562928314784, accuracy: 0.8244556114102528\n",
      "Epoch 1/1\n",
      "188/187 [==============================] - 25s - loss: 0.4767 - acc: 0.8404    \n",
      "epoch: 2, loss: 0.5103619897784899, accuracy: 0.8328308207904873\n",
      "Epoch 1/1\n",
      "188/187 [==============================] - 25s - loss: 0.4569 - acc: 0.8564    \n",
      "epoch: 3, loss: 0.5124339959729257, accuracy: 0.8294807370583616\n",
      "Epoch 1/1\n",
      "188/187 [==============================] - 25s - loss: 0.4201 - acc: 0.8710    \n",
      "epoch: 4, loss: 0.5103467443879924, accuracy: 0.8355108877921623\n",
      "Epoch 1/1\n",
      "188/187 [==============================] - 25s - loss: 0.3663 - acc: 0.8956    \n",
      "epoch: 5, loss: 0.5447332611834783, accuracy: 0.8288107202879748\n",
      "Epoch 1/1\n",
      "188/187 [==============================] - 26s - loss: 0.3712 - acc: 0.8670    \n",
      "epoch: 6, loss: 0.5340726169010303, accuracy: 0.8318257956648591\n",
      "Epoch 1/1\n",
      "  6/187 [..............................] - ETA: 192s - loss: 0.3335 - acc: 0.8958"
     ]
    },
    {
     "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-85-d6aa3286e116>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     10\u001b[0m                        \u001b[0mverbose\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     11\u001b[0m                        \u001b[0mworkers\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[0;32m---> 12\u001b[0;31m                        epochs=1)\n\u001b[0m\u001b[1;32m     13\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mi\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;36m1\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[1;32m     14\u001b[0m         \u001b[0mscore\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtestx_rgb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtesty\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mverbose\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/python3.5/dist-packages/keras/legacy/interfaces.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     86\u001b[0m                 warnings.warn('Update your `' + object_name +\n\u001b[1;32m     87\u001b[0m                               '` call to the Keras 2 API: ' + signature, stacklevel=2)\n\u001b[0;32m---> 88\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\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[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     89\u001b[0m         \u001b[0mwrapper\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_support_signature\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgetargspec\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfunc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     90\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mwrapper\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit_generator\u001b[0;34m(self, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_q_size, workers, pickle_safe, initial_epoch)\u001b[0m\n\u001b[1;32m   1875\u001b[0m                     outs = self.train_on_batch(x, y,\n\u001b[1;32m   1876\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-> 1877\u001b[0;31m                                                class_weight=class_weight)\n\u001b[0m\u001b[1;32m   1878\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1879\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/python3.5/dist-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mtrain_on_batch\u001b[0;34m(self, x, y, sample_weight, class_weight)\u001b[0m\n\u001b[1;32m   1619\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   1620\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-> 1621\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   1622\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   1623\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/python3.5/dist-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2101\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   2102\u001b[0m         updated = session.run(self.outputs + [self.updates_op],\n\u001b[0;32m-> 2103\u001b[0;31m                               feed_dict=feed_dict)\n\u001b[0m\u001b[1;32m   2104\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   2105\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    776\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    777\u001b[0m       result = self._run(None, fetches, feed_dict, options_ptr,\n\u001b[0;32m--> 778\u001b[0;31m                          run_metadata_ptr)\n\u001b[0m\u001b[1;32m    779\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    780\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/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_run\u001b[0;34m(self, handle, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    980\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    981\u001b[0m       results = self._do_run(handle, final_targets, final_fetches,\n\u001b[0;32m--> 982\u001b[0;31m                              feed_dict_string, options, run_metadata)\n\u001b[0m\u001b[1;32m    983\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    984\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/python3.5/dist-packages/tensorflow/python/client/session.py\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   1030\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mhandle\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1031\u001b[0m       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,\n\u001b[0;32m-> 1032\u001b[0;31m                            target_list, options, run_metadata)\n\u001b[0m\u001b[1;32m   1033\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1034\u001b[0m       return self._do_call(_prun_fn, self._session, handle, feed_dict,\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_do_call\u001b[0;34m(self, fn, *args)\u001b[0m\n\u001b[1;32m   1037\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   1038\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\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   1040\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   1041\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/python3.5/dist-packages/tensorflow/python/client/session.py\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   1019\u001b[0m         return tf_session.TF_Run(session, options,\n\u001b[1;32m   1020\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-> 1021\u001b[0;31m                                  status, run_metadata)\n\u001b[0m\u001b[1;32m   1022\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1023\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": [
    "\n",
    "batch_size = 8\n",
    "\n",
    "for i in range(40):\n",
    "    steps = len(trainx)/batch_size\n",
    "    # steps = 1\n",
    "    model2.optimizer.lr.assign((1e-4) / 2)\n",
    "    o = model2.fit_generator(image_gen.flow(trainx_rgb, trainy, batch_size=batch_size),\n",
    "                       steps_per_epoch=steps,\n",
    "                       verbose=1,\n",
    "                       workers=2,\n",
    "                       epochs=1)\n",
    "    if i % 1 == 0:\n",
    "        score = model2.evaluate(testx_rgb, testy, verbose=0)\n",
    "        print('epoch: {}, loss: {}, accuracy: {}'.format(i+1, score[0], score[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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
}