CIFAR10-CNN-TransferLearning-VGG16.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "%config InlineBackend.figure_format = 'retina'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import backend as K\n",
    "from keras.applications.vgg16 import VGG16\n",
    "import tensorflow as tf\n",
    "config = tf.ConfigProto()\n",
    "config.gpu_options.allow_growth=True\n",
    "sess = tf.Session(config=config)\n",
    "K.set_session(sess)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "from numpy.random import seed\n",
    "seed(1)\n",
    "from tensorflow import set_random_seed\n",
    "set_random_seed(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Keras modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.utils import to_categorical\n",
    "from keras.callbacks import ModelCheckpoint, EarlyStopping\n",
    "from keras.models import Model,Sequential\n",
    "from keras.layers.core import Dense, Flatten, Activation\n",
    "from keras.layers import Conv2D, AveragePooling2D, BatchNormalization, Dropout, MaxPooling2D, Input,Add"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cargo datos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import cifar10\n",
    "((x, y), (x_test,y_test)) = cifar10.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, x_valid, y_train, y_valid = train_test_split(x, y, test_size = 0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Paso a float"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = x_train.astype('float32')\n",
    "x_valid = x_valid.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "x_train /= 255\n",
    "x_valid /= 255\n",
    "x_test /= 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_categorical = to_categorical(y_train)\n",
    "y_val_categorical = to_categorical(y_valid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Red neuronal: Transfer learning con VGG16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "preproc=VGG16(include_top = False, pooling = None, weights=\"imagenet\",input_shape=(32,32,3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_5 (InputLayer)         (None, 32, 32, 3)         0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 32, 32, 64)        1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 32, 32, 64)        36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 16, 16, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 16, 16, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 8, 8, 256)         295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 4, 4, 256)         0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 4, 4, 512)         1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 2, 2, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 1, 1, 512)         0         \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 14,714,688\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "preproc.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layer in preproc.layers:\n",
    "    layer.trainable=False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_pre_train=preproc.predict(x_train)\n",
    "x_pre_valid=preproc.predict(x_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 1, 512)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_pre_train[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_6 (InputLayer)         (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "flatten_3 (Flatten)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 256)               65792     \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 10)                2570      \n",
      "=================================================================\n",
      "Total params: 199,690\n",
      "Trainable params: 199,690\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "input_preproc=Input(shape=(1,1,512))\n",
    "\n",
    "net=Flatten()(input_preproc)\n",
    "\n",
    "net=Dense(units=256, activation='relu')(net)\n",
    "net=Dropout(0.2)(net)\n",
    "net=Dense(units=256, activation='relu')(net)\n",
    "net=Dropout(0.2)(net)\n",
    "net=Dense(units=10, activation = 'softmax')(net)\n",
    "\n",
    "model=Model(input_preproc,net)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Callbacks\n",
    "checkpoint_file='checkpoint.cifar10-cnn-TL-VGG16.hdf5'\n",
    "earlystop = EarlyStopping(monitor='val_acc', min_delta=0, patience=5, verbose=2)\n",
    "checkpointer = ModelCheckpoint(monitor=\"val_acc\",filepath=checkpoint_file, verbose=1, save_best_only=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(45000, 32, 32, 3)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(45000, 10)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_categorical.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 45000 samples, validate on 5000 samples\n",
      "Epoch 1/100\n",
      "45000/45000 [==============================] - 1s 32us/step - loss: 1.9860 - acc: 0.2889 - val_loss: 1.6808 - val_acc: 0.4142\n",
      "\n",
      "Epoch 00001: val_acc improved from -inf to 0.41420, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 2/100\n",
      "45000/45000 [==============================] - 1s 24us/step - loss: 1.6559 - acc: 0.4131 - val_loss: 1.5276 - val_acc: 0.4538\n",
      "\n",
      "Epoch 00002: val_acc improved from 0.41420 to 0.45380, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 3/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.5352 - acc: 0.4594 - val_loss: 1.4274 - val_acc: 0.4958\n",
      "\n",
      "Epoch 00003: val_acc improved from 0.45380 to 0.49580, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 4/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.4713 - acc: 0.4826 - val_loss: 1.3792 - val_acc: 0.5138\n",
      "\n",
      "Epoch 00004: val_acc improved from 0.49580 to 0.51380, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 5/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.4287 - acc: 0.4996 - val_loss: 1.3541 - val_acc: 0.5160\n",
      "\n",
      "Epoch 00005: val_acc improved from 0.51380 to 0.51600, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 6/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.3943 - acc: 0.5082 - val_loss: 1.3269 - val_acc: 0.5334\n",
      "\n",
      "Epoch 00006: val_acc improved from 0.51600 to 0.53340, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 7/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.3673 - acc: 0.5200 - val_loss: 1.3119 - val_acc: 0.5282\n",
      "\n",
      "Epoch 00007: val_acc did not improve from 0.53340\n",
      "Epoch 8/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.3461 - acc: 0.5281 - val_loss: 1.2921 - val_acc: 0.5404\n",
      "\n",
      "Epoch 00008: val_acc improved from 0.53340 to 0.54040, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 9/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.3275 - acc: 0.5338 - val_loss: 1.2727 - val_acc: 0.5446\n",
      "\n",
      "Epoch 00009: val_acc improved from 0.54040 to 0.54460, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 10/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.3086 - acc: 0.5406 - val_loss: 1.2959 - val_acc: 0.5382\n",
      "\n",
      "Epoch 00010: val_acc did not improve from 0.54460\n",
      "Epoch 11/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.2963 - acc: 0.5426 - val_loss: 1.2356 - val_acc: 0.5580\n",
      "\n",
      "Epoch 00011: val_acc improved from 0.54460 to 0.55800, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 12/100\n",
      "45000/45000 [==============================] - 1s 24us/step - loss: 1.2832 - acc: 0.5487 - val_loss: 1.2402 - val_acc: 0.5596\n",
      "\n",
      "Epoch 00012: val_acc improved from 0.55800 to 0.55960, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 13/100\n",
      "45000/45000 [==============================] - 1s 22us/step - loss: 1.2689 - acc: 0.5545 - val_loss: 1.2648 - val_acc: 0.5492\n",
      "\n",
      "Epoch 00013: val_acc did not improve from 0.55960\n",
      "Epoch 14/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.2583 - acc: 0.5565 - val_loss: 1.2138 - val_acc: 0.5742\n",
      "\n",
      "Epoch 00014: val_acc improved from 0.55960 to 0.57420, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 15/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.2443 - acc: 0.5604 - val_loss: 1.3200 - val_acc: 0.5308\n",
      "\n",
      "Epoch 00015: val_acc did not improve from 0.57420\n",
      "Epoch 16/100\n",
      "45000/45000 [==============================] - 1s 22us/step - loss: 1.2376 - acc: 0.5662 - val_loss: 1.2519 - val_acc: 0.5564\n",
      "\n",
      "Epoch 00016: val_acc did not improve from 0.57420\n",
      "Epoch 17/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.2308 - acc: 0.5685 - val_loss: 1.2136 - val_acc: 0.5680\n",
      "\n",
      "Epoch 00017: val_acc did not improve from 0.57420\n",
      "Epoch 18/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.2212 - acc: 0.5721 - val_loss: 1.1941 - val_acc: 0.5754\n",
      "\n",
      "Epoch 00018: val_acc improved from 0.57420 to 0.57540, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 19/100\n",
      "45000/45000 [==============================] - 1s 22us/step - loss: 1.2129 - acc: 0.5745 - val_loss: 1.2565 - val_acc: 0.5558\n",
      "\n",
      "Epoch 00019: val_acc did not improve from 0.57540\n",
      "Epoch 20/100\n",
      "45000/45000 [==============================] - 1s 22us/step - loss: 1.2028 - acc: 0.5779 - val_loss: 1.1949 - val_acc: 0.5754\n",
      "\n",
      "Epoch 00020: val_acc did not improve from 0.57540\n",
      "Epoch 21/100\n",
      "45000/45000 [==============================] - 1s 22us/step - loss: 1.1981 - acc: 0.5804 - val_loss: 1.2202 - val_acc: 0.5690\n",
      "\n",
      "Epoch 00021: val_acc did not improve from 0.57540\n",
      "Epoch 22/100\n",
      "45000/45000 [==============================] - 1s 22us/step - loss: 1.1902 - acc: 0.5824 - val_loss: 1.1748 - val_acc: 0.5908\n",
      "\n",
      "Epoch 00022: val_acc improved from 0.57540 to 0.59080, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 23/100\n",
      "45000/45000 [==============================] - 1s 27us/step - loss: 1.1879 - acc: 0.5859 - val_loss: 1.3273 - val_acc: 0.5376\n",
      "\n",
      "Epoch 00023: val_acc did not improve from 0.59080\n",
      "Epoch 24/100\n",
      "45000/45000 [==============================] - 1s 27us/step - loss: 1.1756 - acc: 0.5878 - val_loss: 1.1723 - val_acc: 0.5912\n",
      "\n",
      "Epoch 00024: val_acc improved from 0.59080 to 0.59120, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 25/100\n",
      "45000/45000 [==============================] - 1s 26us/step - loss: 1.1713 - acc: 0.5897 - val_loss: 1.2006 - val_acc: 0.5804\n",
      "\n",
      "Epoch 00025: val_acc did not improve from 0.59120\n",
      "Epoch 26/100\n",
      "45000/45000 [==============================] - 1s 24us/step - loss: 1.1659 - acc: 0.5922 - val_loss: 1.1620 - val_acc: 0.5922\n",
      "\n",
      "Epoch 00026: val_acc improved from 0.59120 to 0.59220, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 27/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.1623 - acc: 0.5927 - val_loss: 1.1502 - val_acc: 0.5994\n",
      "\n",
      "Epoch 00027: val_acc improved from 0.59220 to 0.59940, saving model to checkpoint.cifar10-cnn-TL-VGG16.hdf5\n",
      "Epoch 28/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.1539 - acc: 0.5973 - val_loss: 1.2649 - val_acc: 0.5662\n",
      "\n",
      "Epoch 00028: val_acc did not improve from 0.59940\n",
      "Epoch 29/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.1517 - acc: 0.5951 - val_loss: 1.2494 - val_acc: 0.5522\n",
      "\n",
      "Epoch 00029: val_acc did not improve from 0.59940\n",
      "Epoch 30/100\n",
      "45000/45000 [==============================] - 1s 25us/step - loss: 1.1486 - acc: 0.5994 - val_loss: 1.1591 - val_acc: 0.5898\n",
      "\n",
      "Epoch 00030: val_acc did not improve from 0.59940\n",
      "Epoch 31/100\n",
      "45000/45000 [==============================] - 1s 27us/step - loss: 1.1421 - acc: 0.5986 - val_loss: 1.1369 - val_acc: 0.5960\n",
      "\n",
      "Epoch 00031: val_acc did not improve from 0.59940\n",
      "Epoch 32/100\n",
      "45000/45000 [==============================] - 1s 23us/step - loss: 1.1369 - acc: 0.6032 - val_loss: 1.2114 - val_acc: 0.5664\n",
      "\n",
      "Epoch 00032: val_acc did not improve from 0.59940\n",
      "Epoch 00032: early stopping\n"
     ]
    }
   ],
   "source": [
    "batch_size = 64\n",
    "epocs = 100\n",
    "\n",
    "history = model.fit(x_pre_train, \n",
    "        y_train_categorical,\n",
    "        epochs=epocs, batch_size=batch_size, \n",
    "        verbose=1, \n",
    "        validation_data = (x_pre_valid, y_val_categorical),\n",
    "        callbacks=[checkpointer, earlystop],\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 304,
       "width": 1043
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, (ax1, ax2) = plt.subplots(1, 2, sharex=True, figsize=(18,5))\n",
    "ax1.plot(history.history['loss'], label=\"loss\")\n",
    "ax1.plot(history.history['val_loss'], label=\"val_loss\")\n",
    "ax1.legend()\n",
    "\n",
    "ax2.plot(history.history['acc'], label=\"acc\")\n",
    "ax2.plot(history.history['val_acc'], label=\"val_acc\")\n",
    "ax2.legend()\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.load_weights('checkpoint.cifar10-cnn-TL-VGG16.hdf5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_5 (InputLayer)         (None, 32, 32, 3)         0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 32, 32, 64)        1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 32, 32, 64)        36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 16, 16, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 16, 16, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 8, 8, 256)         295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 4, 4, 256)         0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 4, 4, 512)         1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 2, 2, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "model_6 (Model)              (None, 10)                199690    \n",
      "=================================================================\n",
      "Total params: 14,914,378\n",
      "Trainable params: 199,690\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "total_model_output=model(preproc.output)\n",
    "total_model=Model(preproc.input,total_model_output)\n",
    "total_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_5 (InputLayer)         (None, 32, 32, 3)         0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 32, 32, 64)        1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 32, 32, 64)        36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 16, 16, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 16, 16, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 8, 8, 256)         295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 4, 4, 256)         0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 4, 4, 512)         1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 2, 2, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "model_6 (Model)              (None, 10)                199690    \n",
      "=================================================================\n",
      "Total params: 14,914,378\n",
      "Trainable params: 14,914,378\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "for layer in preproc.layers:\n",
    "    layer.trainable=True\n",
    "total_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_model.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])\n",
    "# Callbacks\n",
    "checkpoint_file='checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5'\n",
    "earlystop = EarlyStopping(monitor='val_acc', min_delta=0, patience=5, verbose=2)\n",
    "checkpointer = ModelCheckpoint(monitor=\"val_acc\",filepath=checkpoint_file, verbose=1, save_best_only=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 45000 samples, validate on 5000 samples\n",
      "Epoch 1/100\n",
      "45000/45000 [==============================] - 18s 398us/step - loss: 1.2997 - acc: 0.5509 - val_loss: 1.1618 - val_acc: 0.5988\n",
      "\n",
      "Epoch 00001: val_acc improved from -inf to 0.59880, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 2/100\n",
      "45000/45000 [==============================] - 17s 382us/step - loss: 0.6904 - acc: 0.7665 - val_loss: 2.4028 - val_acc: 0.4134\n",
      "\n",
      "Epoch 00002: val_acc did not improve from 0.59880\n",
      "Epoch 3/100\n",
      "45000/45000 [==============================] - 17s 387us/step - loss: 0.5357 - acc: 0.8188 - val_loss: 0.7492 - val_acc: 0.7570\n",
      "\n",
      "Epoch 00003: val_acc improved from 0.59880 to 0.75700, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 4/100\n",
      "45000/45000 [==============================] - 17s 385us/step - loss: 0.4323 - acc: 0.8538 - val_loss: 0.4778 - val_acc: 0.8402\n",
      "\n",
      "Epoch 00004: val_acc improved from 0.75700 to 0.84020, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 5/100\n",
      "45000/45000 [==============================] - 17s 388us/step - loss: 0.3552 - acc: 0.8788 - val_loss: 0.4601 - val_acc: 0.8430\n",
      "\n",
      "Epoch 00005: val_acc improved from 0.84020 to 0.84300, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 6/100\n",
      "45000/45000 [==============================] - 17s 385us/step - loss: 0.2904 - acc: 0.9016 - val_loss: 0.5350 - val_acc: 0.8358\n",
      "\n",
      "Epoch 00006: val_acc did not improve from 0.84300\n",
      "Epoch 7/100\n",
      "45000/45000 [==============================] - 17s 389us/step - loss: 0.2340 - acc: 0.9208 - val_loss: 0.6701 - val_acc: 0.8028\n",
      "\n",
      "Epoch 00007: val_acc did not improve from 0.84300\n",
      "Epoch 8/100\n",
      "45000/45000 [==============================] - 18s 389us/step - loss: 0.1879 - acc: 0.9367 - val_loss: 0.4733 - val_acc: 0.8534\n",
      "\n",
      "Epoch 00008: val_acc improved from 0.84300 to 0.85340, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 9/100\n",
      "45000/45000 [==============================] - 17s 388us/step - loss: 0.1398 - acc: 0.9526 - val_loss: 0.5083 - val_acc: 0.8512\n",
      "\n",
      "Epoch 00009: val_acc did not improve from 0.85340\n",
      "Epoch 10/100\n",
      "45000/45000 [==============================] - 18s 389us/step - loss: 0.1121 - acc: 0.9636 - val_loss: 0.6268 - val_acc: 0.8318\n",
      "\n",
      "Epoch 00010: val_acc did not improve from 0.85340\n",
      "Epoch 11/100\n",
      "45000/45000 [==============================] - 18s 389us/step - loss: 0.0865 - acc: 0.9723 - val_loss: 0.5432 - val_acc: 0.8602\n",
      "\n",
      "Epoch 00011: val_acc improved from 0.85340 to 0.86020, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 12/100\n",
      "45000/45000 [==============================] - 17s 388us/step - loss: 0.0648 - acc: 0.9792 - val_loss: 4.1915 - val_acc: 0.4884\n",
      "\n",
      "Epoch 00012: val_acc did not improve from 0.86020\n",
      "Epoch 13/100\n",
      "45000/45000 [==============================] - 18s 392us/step - loss: 0.1790 - acc: 0.9579 - val_loss: 0.5668 - val_acc: 0.8602\n",
      "\n",
      "Epoch 00013: val_acc did not improve from 0.86020\n",
      "Epoch 14/100\n",
      "45000/45000 [==============================] - 18s 390us/step - loss: 0.0441 - acc: 0.9861 - val_loss: 0.5989 - val_acc: 0.8654\n",
      "\n",
      "Epoch 00014: val_acc improved from 0.86020 to 0.86540, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 15/100\n",
      "45000/45000 [==============================] - 18s 395us/step - loss: 0.0331 - acc: 0.9904 - val_loss: 0.6061 - val_acc: 0.8646\n",
      "\n",
      "Epoch 00015: val_acc did not improve from 0.86540\n",
      "Epoch 16/100\n",
      "45000/45000 [==============================] - 18s 396us/step - loss: 0.0272 - acc: 0.9920 - val_loss: 0.6051 - val_acc: 0.8642\n",
      "\n",
      "Epoch 00016: val_acc did not improve from 0.86540\n",
      "Epoch 17/100\n",
      "45000/45000 [==============================] - 18s 396us/step - loss: 0.0349 - acc: 0.9908 - val_loss: 0.5991 - val_acc: 0.8660\n",
      "\n",
      "Epoch 00017: val_acc improved from 0.86540 to 0.86600, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 18/100\n",
      "45000/45000 [==============================] - 17s 388us/step - loss: 0.0065 - acc: 0.9985 - val_loss: 0.6244 - val_acc: 0.8736\n",
      "\n",
      "Epoch 00018: val_acc improved from 0.86600 to 0.87360, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 19/100\n",
      "45000/45000 [==============================] - 18s 391us/step - loss: 0.0327 - acc: 0.9942 - val_loss: 0.5539 - val_acc: 0.8504\n",
      "\n",
      "Epoch 00019: val_acc did not improve from 0.87360\n",
      "Epoch 20/100\n",
      "45000/45000 [==============================] - 18s 392us/step - loss: 0.0164 - acc: 0.9957 - val_loss: 0.6591 - val_acc: 0.8654\n",
      "\n",
      "Epoch 00020: val_acc did not improve from 0.87360\n",
      "Epoch 21/100\n",
      "45000/45000 [==============================] - 18s 394us/step - loss: 0.0052 - acc: 0.9988 - val_loss: 0.6546 - val_acc: 0.8694\n",
      "\n",
      "Epoch 00021: val_acc did not improve from 0.87360\n",
      "Epoch 22/100\n",
      "45000/45000 [==============================] - 18s 394us/step - loss: 0.0020 - acc: 0.9997 - val_loss: 0.6634 - val_acc: 0.8744\n",
      "\n",
      "Epoch 00022: val_acc improved from 0.87360 to 0.87440, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 23/100\n",
      "45000/45000 [==============================] - 18s 392us/step - loss: 5.4294e-04 - acc: 1.0000 - val_loss: 0.6849 - val_acc: 0.8738\n",
      "\n",
      "Epoch 00023: val_acc did not improve from 0.87440\n",
      "Epoch 24/100\n",
      "45000/45000 [==============================] - 18s 394us/step - loss: 3.9031e-04 - acc: 1.0000 - val_loss: 0.7019 - val_acc: 0.8726\n",
      "\n",
      "Epoch 00024: val_acc did not improve from 0.87440\n",
      "Epoch 25/100\n",
      "45000/45000 [==============================] - 18s 393us/step - loss: 2.8815e-04 - acc: 1.0000 - val_loss: 0.7168 - val_acc: 0.8732\n",
      "\n",
      "Epoch 00025: val_acc did not improve from 0.87440\n",
      "Epoch 26/100\n",
      "45000/45000 [==============================] - 18s 396us/step - loss: 2.5745e-04 - acc: 1.0000 - val_loss: 0.7301 - val_acc: 0.8750\n",
      "\n",
      "Epoch 00026: val_acc improved from 0.87440 to 0.87500, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 27/100\n",
      "45000/45000 [==============================] - 18s 394us/step - loss: 2.3247e-04 - acc: 1.0000 - val_loss: 0.7355 - val_acc: 0.8736\n",
      "\n",
      "Epoch 00027: val_acc did not improve from 0.87500\n",
      "Epoch 28/100\n",
      "45000/45000 [==============================] - 18s 396us/step - loss: 2.0943e-04 - acc: 1.0000 - val_loss: 0.7461 - val_acc: 0.8748\n",
      "\n",
      "Epoch 00028: val_acc did not improve from 0.87500\n",
      "Epoch 29/100\n",
      "45000/45000 [==============================] - 18s 394us/step - loss: 2.4824e-04 - acc: 1.0000 - val_loss: 0.7572 - val_acc: 0.8756\n",
      "\n",
      "Epoch 00029: val_acc improved from 0.87500 to 0.87560, saving model to checkpoint.cifar10-cnn-TL-VGG16_fine_tunning.hdf5\n",
      "Epoch 30/100\n",
      "45000/45000 [==============================] - 18s 393us/step - loss: 2.1044e-04 - acc: 1.0000 - val_loss: 0.7647 - val_acc: 0.8736\n",
      "\n",
      "Epoch 00030: val_acc did not improve from 0.87560\n",
      "Epoch 31/100\n",
      "45000/45000 [==============================] - 18s 394us/step - loss: 1.4678e-04 - acc: 1.0000 - val_loss: 0.7704 - val_acc: 0.8736\n",
      "\n",
      "Epoch 00031: val_acc did not improve from 0.87560\n",
      "Epoch 32/100\n",
      "45000/45000 [==============================] - 18s 396us/step - loss: 1.4212e-04 - acc: 1.0000 - val_loss: 0.7754 - val_acc: 0.8734\n",
      "\n",
      "Epoch 00032: val_acc did not improve from 0.87560\n",
      "Epoch 33/100\n",
      "45000/45000 [==============================] - 18s 395us/step - loss: 1.5529e-04 - acc: 1.0000 - val_loss: 0.7824 - val_acc: 0.8734\n",
      "\n",
      "Epoch 00033: val_acc did not improve from 0.87560\n",
      "Epoch 34/100\n",
      "45000/45000 [==============================] - 18s 393us/step - loss: 1.0439e-04 - acc: 1.0000 - val_loss: 0.7868 - val_acc: 0.8730\n",
      "\n",
      "Epoch 00034: val_acc did not improve from 0.87560\n",
      "Epoch 00034: early stopping\n"
     ]
    }
   ],
   "source": [
    "batch_size = 64\n",
    "epocs = 100\n",
    "\n",
    "history = total_model.fit(x_train, \n",
    "        y_train_categorical,\n",
    "        epochs=epocs, batch_size=batch_size, \n",
    "        verbose=1, \n",
    "        validation_data = (x_valid, y_val_categorical),\n",
    "        callbacks=[checkpointer, earlystop],\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (tensorflow-gpu)",
   "language": "python",
   "name": "tensorflow-gpu"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}