Part B.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.1.0\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "print(tf.__version__)\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import cv2\n",
    "import os\n",
    "from glob import glob\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"vgg16\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 128, 128, 3)]     0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 128, 128, 64)      1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 128, 128, 64)      36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 64, 64, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 64, 64, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 64, 64, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 32, 32, 128)       0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 32, 32, 256)       295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 32, 32, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 32, 32, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 16, 16, 256)       0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 16, 16, 512)       1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 16, 16, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 16, 16, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 8, 8, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 8, 8, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 8, 8, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 8, 8, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 4, 4, 512)         0         \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 14,714,688\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "vggModel = tf.keras.applications.VGG16(weights='imagenet', include_top=False, input_shape=(128, 128, 3))\n",
    "\n",
    "print (vggModel.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1020, 128, 128, 3) (1020,)\n",
      "(340, 128, 128, 3) (340,)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "import h5py\n",
    "\n",
    "def loaddataH5():\n",
    "    with h5py.File('data1.h5','r') as hf:\n",
    "        trainX = np.array(hf.get('trainX'))\n",
    "        trainY=np.array(hf.get('trainY'))\n",
    "        valX=np.array(hf.get('valX'))\n",
    "        valY=np.array(hf.get('valY'))\n",
    "        print(trainX.shape,trainY.shape)\n",
    "        print(valX.shape,valY.shape)\n",
    "    return trainX,trainY,valX,valY\n",
    "\n",
    "trainX,trainY,testX,testY=loaddataH5()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0  6 10 16  6  7  2  3  4  1 14  7  3  0 13  8  1 14 14  0 15  4  0  5\n",
      " 13  1  9  8  3 11  9  4 13 12  8  5  5 10 10  7 11 16  4 16 12 11  3  8\n",
      "  0  7  5 13 14 16  2 12 13  6  3 16  0 13 13  4 10 14  0  8 12  3  9 12\n",
      "  7  0 10  3 11 16  0 11 11  6  1  5  2  9 10 11  9 11 12 16 15 12  2 16\n",
      "  3  1 10  6 13  6  8 11  6  8 13  3  7  7  4  8 10  6  2  7  8  0 16  3\n",
      " 11 13  8 16 12  9  5  3  0 12  1  0 15  7  1  8  2 10  3  5  8  1 13  4\n",
      " 10  8  5 15  1  7 13  3  5 11  1  2 11  2 16  2  7 11  8  4 12  5 13  2\n",
      "  3 11 10 13  9 10 10 11  4  8  0  6 12  8 11  7 16  8 15 15 10  6  9 12\n",
      " 10  8  3 14  2 15 14 12 14  9  6  7  6 12  7  8 14  6  9  1  2  0  4 14\n",
      "  9  6  0  2 12  7 14 13 11 12  2  1 13  0  7 14  4 10 10 14  4 13 16  1\n",
      "  4 10 12  0 14 15 15  4  6  2 13  9  9 10 10 13 11 11  5  2  9 14  1  6\n",
      "  6  8  8  9  8 13  9  8  2 10 14  3  0  9  3 15 16 11  2  1 11  8 10  5\n",
      "  4  7 12  9  3 12 16 15  6  1 12 14 16  2  0 15  8 13 13  3 13  5  5 10\n",
      "  7  8 16  7  7 15  5 14  4  2  3 16 15  5  5 11 11  8  7 16  9  9 14 15\n",
      "  9  5  9 16]\n"
     ]
    }
   ],
   "source": [
    "print(testY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "featuresTrain = vggModel.predict(trainX)\n",
    "trainX_vgg = featuresTrain.reshape(featuresTrain.shape[0], -1)\n",
    "featuresTest = vggModel.predict(testX)\n",
    "testX_vgg = featuresTest.reshape(featuresTest.shape[0], -1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import metrics\n",
    "def logisticreg(trainX, trainY, testX, testY):\n",
    "    model = LogisticRegression(random_state=0)\n",
    "    model.fit(trainX, trainY)\n",
    "    results = model.predict(testX)\n",
    "    print(\"Logistic Regression Accuracy:\")\n",
    "    print (metrics.accuracy_score(results, testY))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "def randomforest(trainX, trainY, testX, testY):\n",
    "    model = RandomForestClassifier(200)\n",
    "    model.fit(trainX, trainY)\n",
    "    # evaluate the model\n",
    "    results = model.predict(testX)\n",
    "    print(\"Random Forest Accuracy:\")\n",
    "    print (metrics.accuracy_score(results, testY))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Accuracy:\n",
      "0.8735294117647059\n",
      "Random Forest Accuracy:\n",
      "0.8117647058823529\n"
     ]
    }
   ],
   "source": [
    "logisticreg(trainX_vgg, trainY, testX_vgg, testY)\n",
    "randomforest(trainX_vgg, trainY, testX_vgg, testY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.5/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
      "87916544/87910968 [==============================] - 144s 2us/step\n",
      "Model: \"inception_v3\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_3 (InputLayer)            [(None, 128, 128, 3) 0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d (Conv2D)                 (None, 63, 63, 32)   864         input_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization (BatchNorma (None, 63, 63, 32)   96          conv2d[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "activation (Activation)         (None, 63, 63, 32)   0           batch_normalization[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)               (None, 61, 61, 32)   9216        activation[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_1 (BatchNor (None, 61, 61, 32)   96          conv2d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_1 (Activation)       (None, 61, 61, 32)   0           batch_normalization_1[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)               (None, 61, 61, 64)   18432       activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_2 (BatchNor (None, 61, 61, 64)   192         conv2d_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_2 (Activation)       (None, 61, 61, 64)   0           batch_normalization_2[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D)    (None, 30, 30, 64)   0           activation_2[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)               (None, 30, 30, 80)   5120        max_pooling2d[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_3 (BatchNor (None, 30, 30, 80)   240         conv2d_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_3 (Activation)       (None, 30, 30, 80)   0           batch_normalization_3[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)               (None, 28, 28, 192)  138240      activation_3[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_4 (BatchNor (None, 28, 28, 192)  576         conv2d_4[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_4 (Activation)       (None, 28, 28, 192)  0           batch_normalization_4[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)  (None, 13, 13, 192)  0           activation_4[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_8 (Conv2D)               (None, 13, 13, 64)   12288       max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_8 (BatchNor (None, 13, 13, 64)   192         conv2d_8[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_8 (Activation)       (None, 13, 13, 64)   0           batch_normalization_8[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_6 (Conv2D)               (None, 13, 13, 48)   9216        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_9 (Conv2D)               (None, 13, 13, 96)   55296       activation_8[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_6 (BatchNor (None, 13, 13, 48)   144         conv2d_6[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_9 (BatchNor (None, 13, 13, 96)   288         conv2d_9[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_6 (Activation)       (None, 13, 13, 48)   0           batch_normalization_6[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_9 (Activation)       (None, 13, 13, 96)   0           batch_normalization_9[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d (AveragePooli (None, 13, 13, 192)  0           max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_5 (Conv2D)               (None, 13, 13, 64)   12288       max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_7 (Conv2D)               (None, 13, 13, 64)   76800       activation_6[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_10 (Conv2D)              (None, 13, 13, 96)   82944       activation_9[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_11 (Conv2D)              (None, 13, 13, 32)   6144        average_pooling2d[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_5 (BatchNor (None, 13, 13, 64)   192         conv2d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_7 (BatchNor (None, 13, 13, 64)   192         conv2d_7[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_10 (BatchNo (None, 13, 13, 96)   288         conv2d_10[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_11 (BatchNo (None, 13, 13, 32)   96          conv2d_11[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_5 (Activation)       (None, 13, 13, 64)   0           batch_normalization_5[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_7 (Activation)       (None, 13, 13, 64)   0           batch_normalization_7[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_10 (Activation)      (None, 13, 13, 96)   0           batch_normalization_10[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_11 (Activation)      (None, 13, 13, 32)   0           batch_normalization_11[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed0 (Concatenate)            (None, 13, 13, 256)  0           activation_5[0][0]               \n",
      "                                                                 activation_7[0][0]               \n",
      "                                                                 activation_10[0][0]              \n",
      "                                                                 activation_11[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_15 (Conv2D)              (None, 13, 13, 64)   16384       mixed0[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_15 (BatchNo (None, 13, 13, 64)   192         conv2d_15[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_15 (Activation)      (None, 13, 13, 64)   0           batch_normalization_15[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_13 (Conv2D)              (None, 13, 13, 48)   12288       mixed0[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_16 (Conv2D)              (None, 13, 13, 96)   55296       activation_15[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_13 (BatchNo (None, 13, 13, 48)   144         conv2d_13[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_16 (BatchNo (None, 13, 13, 96)   288         conv2d_16[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_13 (Activation)      (None, 13, 13, 48)   0           batch_normalization_13[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_16 (Activation)      (None, 13, 13, 96)   0           batch_normalization_16[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_1 (AveragePoo (None, 13, 13, 256)  0           mixed0[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_12 (Conv2D)              (None, 13, 13, 64)   16384       mixed0[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_14 (Conv2D)              (None, 13, 13, 64)   76800       activation_13[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_17 (Conv2D)              (None, 13, 13, 96)   82944       activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_18 (Conv2D)              (None, 13, 13, 64)   16384       average_pooling2d_1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_12 (BatchNo (None, 13, 13, 64)   192         conv2d_12[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_14 (BatchNo (None, 13, 13, 64)   192         conv2d_14[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_17 (BatchNo (None, 13, 13, 96)   288         conv2d_17[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_18 (BatchNo (None, 13, 13, 64)   192         conv2d_18[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_12 (Activation)      (None, 13, 13, 64)   0           batch_normalization_12[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_14 (Activation)      (None, 13, 13, 64)   0           batch_normalization_14[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_17 (Activation)      (None, 13, 13, 96)   0           batch_normalization_17[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_18 (Activation)      (None, 13, 13, 64)   0           batch_normalization_18[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed1 (Concatenate)            (None, 13, 13, 288)  0           activation_12[0][0]              \n",
      "                                                                 activation_14[0][0]              \n",
      "                                                                 activation_17[0][0]              \n",
      "                                                                 activation_18[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_22 (Conv2D)              (None, 13, 13, 64)   18432       mixed1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_22 (BatchNo (None, 13, 13, 64)   192         conv2d_22[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_22 (Activation)      (None, 13, 13, 64)   0           batch_normalization_22[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_20 (Conv2D)              (None, 13, 13, 48)   13824       mixed1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_23 (Conv2D)              (None, 13, 13, 96)   55296       activation_22[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_20 (BatchNo (None, 13, 13, 48)   144         conv2d_20[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_23 (BatchNo (None, 13, 13, 96)   288         conv2d_23[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_20 (Activation)      (None, 13, 13, 48)   0           batch_normalization_20[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_23 (Activation)      (None, 13, 13, 96)   0           batch_normalization_23[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_2 (AveragePoo (None, 13, 13, 288)  0           mixed1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_19 (Conv2D)              (None, 13, 13, 64)   18432       mixed1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_21 (Conv2D)              (None, 13, 13, 64)   76800       activation_20[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_24 (Conv2D)              (None, 13, 13, 96)   82944       activation_23[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_25 (Conv2D)              (None, 13, 13, 64)   18432       average_pooling2d_2[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_19 (BatchNo (None, 13, 13, 64)   192         conv2d_19[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_21 (BatchNo (None, 13, 13, 64)   192         conv2d_21[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_24 (BatchNo (None, 13, 13, 96)   288         conv2d_24[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_25 (BatchNo (None, 13, 13, 64)   192         conv2d_25[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_19 (Activation)      (None, 13, 13, 64)   0           batch_normalization_19[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_21 (Activation)      (None, 13, 13, 64)   0           batch_normalization_21[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_24 (Activation)      (None, 13, 13, 96)   0           batch_normalization_24[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_25 (Activation)      (None, 13, 13, 64)   0           batch_normalization_25[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed2 (Concatenate)            (None, 13, 13, 288)  0           activation_19[0][0]              \n",
      "                                                                 activation_21[0][0]              \n",
      "                                                                 activation_24[0][0]              \n",
      "                                                                 activation_25[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_27 (Conv2D)              (None, 13, 13, 64)   18432       mixed2[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_27 (BatchNo (None, 13, 13, 64)   192         conv2d_27[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_27 (Activation)      (None, 13, 13, 64)   0           batch_normalization_27[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_28 (Conv2D)              (None, 13, 13, 96)   55296       activation_27[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_28 (BatchNo (None, 13, 13, 96)   288         conv2d_28[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_28 (Activation)      (None, 13, 13, 96)   0           batch_normalization_28[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_26 (Conv2D)              (None, 6, 6, 384)    995328      mixed2[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_29 (Conv2D)              (None, 6, 6, 96)     82944       activation_28[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_26 (BatchNo (None, 6, 6, 384)    1152        conv2d_26[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_29 (BatchNo (None, 6, 6, 96)     288         conv2d_29[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_26 (Activation)      (None, 6, 6, 384)    0           batch_normalization_26[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_29 (Activation)      (None, 6, 6, 96)     0           batch_normalization_29[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2D)  (None, 6, 6, 288)    0           mixed2[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "mixed3 (Concatenate)            (None, 6, 6, 768)    0           activation_26[0][0]              \n",
      "                                                                 activation_29[0][0]              \n",
      "                                                                 max_pooling2d_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_34 (Conv2D)              (None, 6, 6, 128)    98304       mixed3[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_34 (BatchNo (None, 6, 6, 128)    384         conv2d_34[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_34 (Activation)      (None, 6, 6, 128)    0           batch_normalization_34[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_35 (Conv2D)              (None, 6, 6, 128)    114688      activation_34[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_35 (BatchNo (None, 6, 6, 128)    384         conv2d_35[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_35 (Activation)      (None, 6, 6, 128)    0           batch_normalization_35[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_31 (Conv2D)              (None, 6, 6, 128)    98304       mixed3[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_36 (Conv2D)              (None, 6, 6, 128)    114688      activation_35[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_31 (BatchNo (None, 6, 6, 128)    384         conv2d_31[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_36 (BatchNo (None, 6, 6, 128)    384         conv2d_36[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_31 (Activation)      (None, 6, 6, 128)    0           batch_normalization_31[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_36 (Activation)      (None, 6, 6, 128)    0           batch_normalization_36[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_32 (Conv2D)              (None, 6, 6, 128)    114688      activation_31[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_37 (Conv2D)              (None, 6, 6, 128)    114688      activation_36[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_32 (BatchNo (None, 6, 6, 128)    384         conv2d_32[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_37 (BatchNo (None, 6, 6, 128)    384         conv2d_37[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_32 (Activation)      (None, 6, 6, 128)    0           batch_normalization_32[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_37 (Activation)      (None, 6, 6, 128)    0           batch_normalization_37[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_3 (AveragePoo (None, 6, 6, 768)    0           mixed3[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_30 (Conv2D)              (None, 6, 6, 192)    147456      mixed3[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_33 (Conv2D)              (None, 6, 6, 192)    172032      activation_32[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_38 (Conv2D)              (None, 6, 6, 192)    172032      activation_37[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_39 (Conv2D)              (None, 6, 6, 192)    147456      average_pooling2d_3[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_30 (BatchNo (None, 6, 6, 192)    576         conv2d_30[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_33 (BatchNo (None, 6, 6, 192)    576         conv2d_33[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_38 (BatchNo (None, 6, 6, 192)    576         conv2d_38[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_39 (BatchNo (None, 6, 6, 192)    576         conv2d_39[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_30 (Activation)      (None, 6, 6, 192)    0           batch_normalization_30[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_33 (Activation)      (None, 6, 6, 192)    0           batch_normalization_33[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_38 (Activation)      (None, 6, 6, 192)    0           batch_normalization_38[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_39 (Activation)      (None, 6, 6, 192)    0           batch_normalization_39[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed4 (Concatenate)            (None, 6, 6, 768)    0           activation_30[0][0]              \n",
      "                                                                 activation_33[0][0]              \n",
      "                                                                 activation_38[0][0]              \n",
      "                                                                 activation_39[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_44 (Conv2D)              (None, 6, 6, 160)    122880      mixed4[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_44 (BatchNo (None, 6, 6, 160)    480         conv2d_44[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_44 (Activation)      (None, 6, 6, 160)    0           batch_normalization_44[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_45 (Conv2D)              (None, 6, 6, 160)    179200      activation_44[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_45 (BatchNo (None, 6, 6, 160)    480         conv2d_45[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_45 (Activation)      (None, 6, 6, 160)    0           batch_normalization_45[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_41 (Conv2D)              (None, 6, 6, 160)    122880      mixed4[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_46 (Conv2D)              (None, 6, 6, 160)    179200      activation_45[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_41 (BatchNo (None, 6, 6, 160)    480         conv2d_41[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_46 (BatchNo (None, 6, 6, 160)    480         conv2d_46[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_41 (Activation)      (None, 6, 6, 160)    0           batch_normalization_41[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_46 (Activation)      (None, 6, 6, 160)    0           batch_normalization_46[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_42 (Conv2D)              (None, 6, 6, 160)    179200      activation_41[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_47 (Conv2D)              (None, 6, 6, 160)    179200      activation_46[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_42 (BatchNo (None, 6, 6, 160)    480         conv2d_42[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_47 (BatchNo (None, 6, 6, 160)    480         conv2d_47[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_42 (Activation)      (None, 6, 6, 160)    0           batch_normalization_42[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_47 (Activation)      (None, 6, 6, 160)    0           batch_normalization_47[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_4 (AveragePoo (None, 6, 6, 768)    0           mixed4[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_40 (Conv2D)              (None, 6, 6, 192)    147456      mixed4[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_43 (Conv2D)              (None, 6, 6, 192)    215040      activation_42[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_48 (Conv2D)              (None, 6, 6, 192)    215040      activation_47[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_49 (Conv2D)              (None, 6, 6, 192)    147456      average_pooling2d_4[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_40 (BatchNo (None, 6, 6, 192)    576         conv2d_40[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_43 (BatchNo (None, 6, 6, 192)    576         conv2d_43[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_48 (BatchNo (None, 6, 6, 192)    576         conv2d_48[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_49 (BatchNo (None, 6, 6, 192)    576         conv2d_49[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_40 (Activation)      (None, 6, 6, 192)    0           batch_normalization_40[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_43 (Activation)      (None, 6, 6, 192)    0           batch_normalization_43[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_48 (Activation)      (None, 6, 6, 192)    0           batch_normalization_48[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_49 (Activation)      (None, 6, 6, 192)    0           batch_normalization_49[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed5 (Concatenate)            (None, 6, 6, 768)    0           activation_40[0][0]              \n",
      "                                                                 activation_43[0][0]              \n",
      "                                                                 activation_48[0][0]              \n",
      "                                                                 activation_49[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_54 (Conv2D)              (None, 6, 6, 160)    122880      mixed5[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_54 (BatchNo (None, 6, 6, 160)    480         conv2d_54[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_54 (Activation)      (None, 6, 6, 160)    0           batch_normalization_54[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_55 (Conv2D)              (None, 6, 6, 160)    179200      activation_54[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_55 (BatchNo (None, 6, 6, 160)    480         conv2d_55[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_55 (Activation)      (None, 6, 6, 160)    0           batch_normalization_55[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_51 (Conv2D)              (None, 6, 6, 160)    122880      mixed5[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_56 (Conv2D)              (None, 6, 6, 160)    179200      activation_55[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_51 (BatchNo (None, 6, 6, 160)    480         conv2d_51[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_56 (BatchNo (None, 6, 6, 160)    480         conv2d_56[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_51 (Activation)      (None, 6, 6, 160)    0           batch_normalization_51[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_56 (Activation)      (None, 6, 6, 160)    0           batch_normalization_56[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_52 (Conv2D)              (None, 6, 6, 160)    179200      activation_51[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_57 (Conv2D)              (None, 6, 6, 160)    179200      activation_56[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_52 (BatchNo (None, 6, 6, 160)    480         conv2d_52[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_57 (BatchNo (None, 6, 6, 160)    480         conv2d_57[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_52 (Activation)      (None, 6, 6, 160)    0           batch_normalization_52[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_57 (Activation)      (None, 6, 6, 160)    0           batch_normalization_57[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_5 (AveragePoo (None, 6, 6, 768)    0           mixed5[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_50 (Conv2D)              (None, 6, 6, 192)    147456      mixed5[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_53 (Conv2D)              (None, 6, 6, 192)    215040      activation_52[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_58 (Conv2D)              (None, 6, 6, 192)    215040      activation_57[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_59 (Conv2D)              (None, 6, 6, 192)    147456      average_pooling2d_5[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_50 (BatchNo (None, 6, 6, 192)    576         conv2d_50[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_53 (BatchNo (None, 6, 6, 192)    576         conv2d_53[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_58 (BatchNo (None, 6, 6, 192)    576         conv2d_58[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_59 (BatchNo (None, 6, 6, 192)    576         conv2d_59[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_50 (Activation)      (None, 6, 6, 192)    0           batch_normalization_50[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_53 (Activation)      (None, 6, 6, 192)    0           batch_normalization_53[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_58 (Activation)      (None, 6, 6, 192)    0           batch_normalization_58[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_59 (Activation)      (None, 6, 6, 192)    0           batch_normalization_59[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed6 (Concatenate)            (None, 6, 6, 768)    0           activation_50[0][0]              \n",
      "                                                                 activation_53[0][0]              \n",
      "                                                                 activation_58[0][0]              \n",
      "                                                                 activation_59[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_64 (Conv2D)              (None, 6, 6, 192)    147456      mixed6[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_64 (BatchNo (None, 6, 6, 192)    576         conv2d_64[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_64 (Activation)      (None, 6, 6, 192)    0           batch_normalization_64[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_65 (Conv2D)              (None, 6, 6, 192)    258048      activation_64[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_65 (BatchNo (None, 6, 6, 192)    576         conv2d_65[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_65 (Activation)      (None, 6, 6, 192)    0           batch_normalization_65[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_61 (Conv2D)              (None, 6, 6, 192)    147456      mixed6[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_66 (Conv2D)              (None, 6, 6, 192)    258048      activation_65[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_61 (BatchNo (None, 6, 6, 192)    576         conv2d_61[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_66 (BatchNo (None, 6, 6, 192)    576         conv2d_66[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_61 (Activation)      (None, 6, 6, 192)    0           batch_normalization_61[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_66 (Activation)      (None, 6, 6, 192)    0           batch_normalization_66[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_62 (Conv2D)              (None, 6, 6, 192)    258048      activation_61[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_67 (Conv2D)              (None, 6, 6, 192)    258048      activation_66[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_62 (BatchNo (None, 6, 6, 192)    576         conv2d_62[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_67 (BatchNo (None, 6, 6, 192)    576         conv2d_67[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_62 (Activation)      (None, 6, 6, 192)    0           batch_normalization_62[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_67 (Activation)      (None, 6, 6, 192)    0           batch_normalization_67[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_6 (AveragePoo (None, 6, 6, 768)    0           mixed6[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_60 (Conv2D)              (None, 6, 6, 192)    147456      mixed6[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_63 (Conv2D)              (None, 6, 6, 192)    258048      activation_62[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_68 (Conv2D)              (None, 6, 6, 192)    258048      activation_67[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_69 (Conv2D)              (None, 6, 6, 192)    147456      average_pooling2d_6[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_60 (BatchNo (None, 6, 6, 192)    576         conv2d_60[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_63 (BatchNo (None, 6, 6, 192)    576         conv2d_63[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_68 (BatchNo (None, 6, 6, 192)    576         conv2d_68[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_69 (BatchNo (None, 6, 6, 192)    576         conv2d_69[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_60 (Activation)      (None, 6, 6, 192)    0           batch_normalization_60[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_63 (Activation)      (None, 6, 6, 192)    0           batch_normalization_63[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_68 (Activation)      (None, 6, 6, 192)    0           batch_normalization_68[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_69 (Activation)      (None, 6, 6, 192)    0           batch_normalization_69[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed7 (Concatenate)            (None, 6, 6, 768)    0           activation_60[0][0]              \n",
      "                                                                 activation_63[0][0]              \n",
      "                                                                 activation_68[0][0]              \n",
      "                                                                 activation_69[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_72 (Conv2D)              (None, 6, 6, 192)    147456      mixed7[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_72 (BatchNo (None, 6, 6, 192)    576         conv2d_72[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_72 (Activation)      (None, 6, 6, 192)    0           batch_normalization_72[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_73 (Conv2D)              (None, 6, 6, 192)    258048      activation_72[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_73 (BatchNo (None, 6, 6, 192)    576         conv2d_73[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_73 (Activation)      (None, 6, 6, 192)    0           batch_normalization_73[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_70 (Conv2D)              (None, 6, 6, 192)    147456      mixed7[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_74 (Conv2D)              (None, 6, 6, 192)    258048      activation_73[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_70 (BatchNo (None, 6, 6, 192)    576         conv2d_70[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_74 (BatchNo (None, 6, 6, 192)    576         conv2d_74[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_70 (Activation)      (None, 6, 6, 192)    0           batch_normalization_70[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_74 (Activation)      (None, 6, 6, 192)    0           batch_normalization_74[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_71 (Conv2D)              (None, 2, 2, 320)    552960      activation_70[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_75 (Conv2D)              (None, 2, 2, 192)    331776      activation_74[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_71 (BatchNo (None, 2, 2, 320)    960         conv2d_71[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_75 (BatchNo (None, 2, 2, 192)    576         conv2d_75[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_71 (Activation)      (None, 2, 2, 320)    0           batch_normalization_71[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_75 (Activation)      (None, 2, 2, 192)    0           batch_normalization_75[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2D)  (None, 2, 2, 768)    0           mixed7[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "mixed8 (Concatenate)            (None, 2, 2, 1280)   0           activation_71[0][0]              \n",
      "                                                                 activation_75[0][0]              \n",
      "                                                                 max_pooling2d_3[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_80 (Conv2D)              (None, 2, 2, 448)    573440      mixed8[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_80 (BatchNo (None, 2, 2, 448)    1344        conv2d_80[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_80 (Activation)      (None, 2, 2, 448)    0           batch_normalization_80[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_77 (Conv2D)              (None, 2, 2, 384)    491520      mixed8[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_81 (Conv2D)              (None, 2, 2, 384)    1548288     activation_80[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_77 (BatchNo (None, 2, 2, 384)    1152        conv2d_77[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_81 (BatchNo (None, 2, 2, 384)    1152        conv2d_81[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_77 (Activation)      (None, 2, 2, 384)    0           batch_normalization_77[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_81 (Activation)      (None, 2, 2, 384)    0           batch_normalization_81[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_78 (Conv2D)              (None, 2, 2, 384)    442368      activation_77[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_79 (Conv2D)              (None, 2, 2, 384)    442368      activation_77[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_82 (Conv2D)              (None, 2, 2, 384)    442368      activation_81[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_83 (Conv2D)              (None, 2, 2, 384)    442368      activation_81[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_7 (AveragePoo (None, 2, 2, 1280)   0           mixed8[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_76 (Conv2D)              (None, 2, 2, 320)    409600      mixed8[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_78 (BatchNo (None, 2, 2, 384)    1152        conv2d_78[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_79 (BatchNo (None, 2, 2, 384)    1152        conv2d_79[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_82 (BatchNo (None, 2, 2, 384)    1152        conv2d_82[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_83 (BatchNo (None, 2, 2, 384)    1152        conv2d_83[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_84 (Conv2D)              (None, 2, 2, 192)    245760      average_pooling2d_7[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_76 (BatchNo (None, 2, 2, 320)    960         conv2d_76[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_78 (Activation)      (None, 2, 2, 384)    0           batch_normalization_78[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_79 (Activation)      (None, 2, 2, 384)    0           batch_normalization_79[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_82 (Activation)      (None, 2, 2, 384)    0           batch_normalization_82[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_83 (Activation)      (None, 2, 2, 384)    0           batch_normalization_83[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_84 (BatchNo (None, 2, 2, 192)    576         conv2d_84[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_76 (Activation)      (None, 2, 2, 320)    0           batch_normalization_76[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed9_0 (Concatenate)          (None, 2, 2, 768)    0           activation_78[0][0]              \n",
      "                                                                 activation_79[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "concatenate (Concatenate)       (None, 2, 2, 768)    0           activation_82[0][0]              \n",
      "                                                                 activation_83[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_84 (Activation)      (None, 2, 2, 192)    0           batch_normalization_84[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed9 (Concatenate)            (None, 2, 2, 2048)   0           activation_76[0][0]              \n",
      "                                                                 mixed9_0[0][0]                   \n",
      "                                                                 concatenate[0][0]                \n",
      "                                                                 activation_84[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_89 (Conv2D)              (None, 2, 2, 448)    917504      mixed9[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_89 (BatchNo (None, 2, 2, 448)    1344        conv2d_89[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_89 (Activation)      (None, 2, 2, 448)    0           batch_normalization_89[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_86 (Conv2D)              (None, 2, 2, 384)    786432      mixed9[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_90 (Conv2D)              (None, 2, 2, 384)    1548288     activation_89[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_86 (BatchNo (None, 2, 2, 384)    1152        conv2d_86[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_90 (BatchNo (None, 2, 2, 384)    1152        conv2d_90[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_86 (Activation)      (None, 2, 2, 384)    0           batch_normalization_86[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_90 (Activation)      (None, 2, 2, 384)    0           batch_normalization_90[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_87 (Conv2D)              (None, 2, 2, 384)    442368      activation_86[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_88 (Conv2D)              (None, 2, 2, 384)    442368      activation_86[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_91 (Conv2D)              (None, 2, 2, 384)    442368      activation_90[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_92 (Conv2D)              (None, 2, 2, 384)    442368      activation_90[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_8 (AveragePoo (None, 2, 2, 2048)   0           mixed9[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_85 (Conv2D)              (None, 2, 2, 320)    655360      mixed9[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_87 (BatchNo (None, 2, 2, 384)    1152        conv2d_87[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_88 (BatchNo (None, 2, 2, 384)    1152        conv2d_88[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_91 (BatchNo (None, 2, 2, 384)    1152        conv2d_91[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_92 (BatchNo (None, 2, 2, 384)    1152        conv2d_92[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_93 (Conv2D)              (None, 2, 2, 192)    393216      average_pooling2d_8[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_85 (BatchNo (None, 2, 2, 320)    960         conv2d_85[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_87 (Activation)      (None, 2, 2, 384)    0           batch_normalization_87[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_88 (Activation)      (None, 2, 2, 384)    0           batch_normalization_88[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_91 (Activation)      (None, 2, 2, 384)    0           batch_normalization_91[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_92 (Activation)      (None, 2, 2, 384)    0           batch_normalization_92[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_93 (BatchNo (None, 2, 2, 192)    576         conv2d_93[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_85 (Activation)      (None, 2, 2, 320)    0           batch_normalization_85[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed9_1 (Concatenate)          (None, 2, 2, 768)    0           activation_87[0][0]              \n",
      "                                                                 activation_88[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 2, 2, 768)    0           activation_91[0][0]              \n",
      "                                                                 activation_92[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_93 (Activation)      (None, 2, 2, 192)    0           batch_normalization_93[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "mixed10 (Concatenate)           (None, 2, 2, 2048)   0           activation_85[0][0]              \n",
      "                                                                 mixed9_1[0][0]                   \n",
      "                                                                 concatenate_1[0][0]              \n",
      "                                                                 activation_93[0][0]              \n",
      "==================================================================================================\n",
      "Total params: 21,802,784\n",
      "Trainable params: 21,768,352\n",
      "Non-trainable params: 34,432\n",
      "__________________________________________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "incepmodel = tf.keras.applications.InceptionV3(weights='imagenet', include_top=False, input_shape=(128, 128, 3))\n",
    "\n",
    "print (incepmodel.summary())\n",
    "\n",
    "featuresTrain = incepmodel.predict(trainX)\n",
    "trainX_incep = featuresTrain.reshape(featuresTrain.shape[0], -1)\n",
    "featuresTest = incepmodel.predict(testX)\n",
    "testX_incep = featuresTest.reshape(featuresTest.shape[0], -1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Accuracy:\n",
      "0.85\n",
      "Random Forest Accuracy:\n",
      "0.8147058823529412\n"
     ]
    }
   ],
   "source": [
    "logisticreg(trainX_incep, trainY, testX_incep, testY)\n",
    "randomforest(trainX_incep, trainY, testX_incep, testY)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fine tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1020 samples, validate on 340 samples\n",
      "Epoch 1/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 2.7615 - accuracy: 0.1167 - val_loss: 2.4394 - val_accuracy: 0.3000\n",
      "Epoch 2/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 2.4667 - accuracy: 0.2196 - val_loss: 2.1031 - val_accuracy: 0.3382\n",
      "Epoch 3/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 2.1703 - accuracy: 0.3108 - val_loss: 1.8082 - val_accuracy: 0.5118\n",
      "Epoch 4/100\n",
      "1020/1020 [==============================] - 5s 4ms/sample - loss: 1.9237 - accuracy: 0.3941 - val_loss: 1.5419 - val_accuracy: 0.6000\n",
      "Epoch 5/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 1.6580 - accuracy: 0.4922 - val_loss: 1.5104 - val_accuracy: 0.4912\n",
      "Epoch 6/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 1.5607 - accuracy: 0.5176 - val_loss: 1.3365 - val_accuracy: 0.5824\n",
      "Epoch 7/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 1.3747 - accuracy: 0.5775 - val_loss: 1.4465 - val_accuracy: 0.5529\n",
      "Epoch 8/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 1.2305 - accuracy: 0.6157 - val_loss: 0.9229 - val_accuracy: 0.7529\n",
      "Epoch 9/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 1.1036 - accuracy: 0.6539 - val_loss: 0.9447 - val_accuracy: 0.7206\n",
      "Epoch 10/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.9987 - accuracy: 0.6990 - val_loss: 0.8450 - val_accuracy: 0.7676\n",
      "Epoch 11/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.9547 - accuracy: 0.7078 - val_loss: 0.7785 - val_accuracy: 0.7941\n",
      "Epoch 12/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.8097 - accuracy: 0.7627 - val_loss: 0.6810 - val_accuracy: 0.8147\n",
      "Epoch 13/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.8151 - accuracy: 0.7441 - val_loss: 0.7410 - val_accuracy: 0.7676\n",
      "Epoch 14/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.7114 - accuracy: 0.7980 - val_loss: 0.6780 - val_accuracy: 0.8029\n",
      "Epoch 15/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.6731 - accuracy: 0.7951 - val_loss: 0.6185 - val_accuracy: 0.8235\n",
      "Epoch 16/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.5945 - accuracy: 0.8157 - val_loss: 0.6067 - val_accuracy: 0.8235\n",
      "Epoch 17/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.5543 - accuracy: 0.8441 - val_loss: 0.7253 - val_accuracy: 0.7794\n",
      "Epoch 18/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.5220 - accuracy: 0.8500 - val_loss: 0.5474 - val_accuracy: 0.8324\n",
      "Epoch 19/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.4775 - accuracy: 0.8647 - val_loss: 0.6037 - val_accuracy: 0.7971\n",
      "Epoch 20/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.4622 - accuracy: 0.8608 - val_loss: 0.5473 - val_accuracy: 0.8559\n",
      "Epoch 21/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.4141 - accuracy: 0.8833 - val_loss: 0.5718 - val_accuracy: 0.8176\n",
      "Epoch 22/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.4000 - accuracy: 0.8882 - val_loss: 0.7555 - val_accuracy: 0.7471\n",
      "Epoch 23/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.3674 - accuracy: 0.9010 - val_loss: 0.5153 - val_accuracy: 0.8382\n",
      "Epoch 24/100\n",
      "1020/1020 [==============================] - 5s 4ms/sample - loss: 0.3456 - accuracy: 0.9078 - val_loss: 0.5047 - val_accuracy: 0.8471\n",
      "Epoch 25/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.3304 - accuracy: 0.9167 - val_loss: 0.4846 - val_accuracy: 0.8529\n",
      "Epoch 26/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.3138 - accuracy: 0.9176 - val_loss: 0.4957 - val_accuracy: 0.8382\n",
      "Epoch 27/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.2987 - accuracy: 0.9196 - val_loss: 0.4468 - val_accuracy: 0.8559\n",
      "Epoch 28/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.3074 - accuracy: 0.9275 - val_loss: 0.5096 - val_accuracy: 0.8353\n",
      "Epoch 29/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.2727 - accuracy: 0.9245 - val_loss: 0.4642 - val_accuracy: 0.8529\n",
      "Epoch 30/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.2443 - accuracy: 0.9431 - val_loss: 0.5285 - val_accuracy: 0.8265\n",
      "Epoch 31/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.2201 - accuracy: 0.9500 - val_loss: 0.4626 - val_accuracy: 0.8412\n",
      "Epoch 32/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.2173 - accuracy: 0.9510 - val_loss: 0.4803 - val_accuracy: 0.8471\n",
      "Epoch 33/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.1968 - accuracy: 0.9667 - val_loss: 0.4173 - val_accuracy: 0.8559\n",
      "Epoch 34/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1915 - accuracy: 0.9569 - val_loss: 0.4118 - val_accuracy: 0.8559\n",
      "Epoch 35/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1800 - accuracy: 0.9510 - val_loss: 0.4896 - val_accuracy: 0.8559\n",
      "Epoch 36/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1673 - accuracy: 0.9676 - val_loss: 0.4707 - val_accuracy: 0.8647\n",
      "Epoch 37/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1567 - accuracy: 0.9696 - val_loss: 0.4818 - val_accuracy: 0.8500\n",
      "Epoch 38/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1453 - accuracy: 0.9745 - val_loss: 0.4264 - val_accuracy: 0.8412\n",
      "Epoch 39/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1391 - accuracy: 0.9706 - val_loss: 0.5014 - val_accuracy: 0.8471\n",
      "Epoch 40/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1549 - accuracy: 0.9657 - val_loss: 0.5243 - val_accuracy: 0.8324\n",
      "Epoch 41/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1290 - accuracy: 0.9725 - val_loss: 0.4394 - val_accuracy: 0.8559\n",
      "Epoch 42/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.1344 - accuracy: 0.9725 - val_loss: 0.3971 - val_accuracy: 0.8647\n",
      "Epoch 43/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1130 - accuracy: 0.9804 - val_loss: 0.4096 - val_accuracy: 0.8706\n",
      "Epoch 44/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1037 - accuracy: 0.9804 - val_loss: 0.4548 - val_accuracy: 0.8500\n",
      "Epoch 45/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1056 - accuracy: 0.9824 - val_loss: 0.4241 - val_accuracy: 0.8500\n",
      "Epoch 46/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1021 - accuracy: 0.9853 - val_loss: 0.4029 - val_accuracy: 0.8529\n",
      "Epoch 47/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1113 - accuracy: 0.9765 - val_loss: 0.4981 - val_accuracy: 0.8382\n",
      "Epoch 48/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1045 - accuracy: 0.9804 - val_loss: 0.4288 - val_accuracy: 0.8500\n",
      "Epoch 49/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0988 - accuracy: 0.9863 - val_loss: 0.4206 - val_accuracy: 0.8588\n",
      "Epoch 50/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.1001 - accuracy: 0.9794 - val_loss: 0.4248 - val_accuracy: 0.8559\n",
      "Epoch 51/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0791 - accuracy: 0.9902 - val_loss: 0.4405 - val_accuracy: 0.8647\n",
      "Epoch 52/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0935 - accuracy: 0.9853 - val_loss: 0.4205 - val_accuracy: 0.8500\n",
      "Epoch 53/100\n",
      "1020/1020 [==============================] - 5s 4ms/sample - loss: 0.0940 - accuracy: 0.9853 - val_loss: 0.3811 - val_accuracy: 0.8676\n",
      "Epoch 54/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0831 - accuracy: 0.9843 - val_loss: 0.4449 - val_accuracy: 0.8529\n",
      "Epoch 55/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0637 - accuracy: 0.9892 - val_loss: 0.4014 - val_accuracy: 0.8529\n",
      "Epoch 56/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0795 - accuracy: 0.9863 - val_loss: 0.3871 - val_accuracy: 0.8559\n",
      "Epoch 57/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0772 - accuracy: 0.9902 - val_loss: 0.4196 - val_accuracy: 0.8471\n",
      "Epoch 58/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0679 - accuracy: 0.9912 - val_loss: 0.4004 - val_accuracy: 0.8676\n",
      "Epoch 59/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0705 - accuracy: 0.9941 - val_loss: 0.3867 - val_accuracy: 0.8588\n",
      "Epoch 60/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0633 - accuracy: 0.9922 - val_loss: 0.3855 - val_accuracy: 0.8559\n",
      "Epoch 61/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0627 - accuracy: 0.9902 - val_loss: 0.4099 - val_accuracy: 0.8529\n",
      "Epoch 62/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0608 - accuracy: 0.9931 - val_loss: 0.3914 - val_accuracy: 0.8618\n",
      "Epoch 63/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0557 - accuracy: 0.9961 - val_loss: 0.4337 - val_accuracy: 0.8529\n",
      "Epoch 64/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0555 - accuracy: 0.9912 - val_loss: 0.3888 - val_accuracy: 0.8559\n",
      "Epoch 65/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0483 - accuracy: 0.9971 - val_loss: 0.4073 - val_accuracy: 0.8559\n",
      "Epoch 66/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0652 - accuracy: 0.9892 - val_loss: 0.4079 - val_accuracy: 0.8529\n",
      "Epoch 67/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0526 - accuracy: 0.9912 - val_loss: 0.3846 - val_accuracy: 0.8676\n",
      "Epoch 68/100\n",
      "1020/1020 [==============================] - 5s 5ms/sample - loss: 0.0476 - accuracy: 0.9951 - val_loss: 0.3757 - val_accuracy: 0.8588\n",
      "Epoch 69/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0512 - accuracy: 0.9941 - val_loss: 0.3870 - val_accuracy: 0.8618\n",
      "Epoch 70/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0448 - accuracy: 0.9912 - val_loss: 0.4019 - val_accuracy: 0.8588\n",
      "Epoch 71/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0473 - accuracy: 0.9922 - val_loss: 0.4131 - val_accuracy: 0.8559\n",
      "Epoch 72/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0492 - accuracy: 0.9941 - val_loss: 0.3826 - val_accuracy: 0.8588\n",
      "Epoch 73/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0468 - accuracy: 0.9951 - val_loss: 0.3944 - val_accuracy: 0.8559\n",
      "Epoch 74/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0462 - accuracy: 0.9951 - val_loss: 0.3885 - val_accuracy: 0.8588\n",
      "Epoch 75/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0424 - accuracy: 0.9961 - val_loss: 0.3841 - val_accuracy: 0.8618\n",
      "Epoch 76/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0358 - accuracy: 0.9980 - val_loss: 0.3785 - val_accuracy: 0.8529\n",
      "Epoch 77/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0500 - accuracy: 0.9922 - val_loss: 0.3721 - val_accuracy: 0.8588\n",
      "Epoch 78/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0391 - accuracy: 0.9971 - val_loss: 0.3893 - val_accuracy: 0.8647\n",
      "Epoch 79/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0394 - accuracy: 0.9951 - val_loss: 0.4167 - val_accuracy: 0.8529\n",
      "Epoch 80/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0404 - accuracy: 1.0000 - val_loss: 0.3749 - val_accuracy: 0.8529\n",
      "Epoch 81/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0406 - accuracy: 0.9941 - val_loss: 0.3853 - val_accuracy: 0.8588\n",
      "Epoch 82/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0387 - accuracy: 0.9941 - val_loss: 0.3880 - val_accuracy: 0.8529\n",
      "Epoch 83/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0343 - accuracy: 0.9980 - val_loss: 0.3784 - val_accuracy: 0.8559\n",
      "Epoch 84/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0429 - accuracy: 0.9951 - val_loss: 0.4245 - val_accuracy: 0.8588\n",
      "Epoch 85/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0371 - accuracy: 0.9961 - val_loss: 0.3797 - val_accuracy: 0.8647\n",
      "Epoch 86/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0294 - accuracy: 0.9990 - val_loss: 0.4017 - val_accuracy: 0.8588\n",
      "Epoch 87/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0297 - accuracy: 0.9980 - val_loss: 0.3849 - val_accuracy: 0.8471\n",
      "Epoch 88/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0347 - accuracy: 0.9971 - val_loss: 0.3892 - val_accuracy: 0.8559\n",
      "Epoch 89/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0355 - accuracy: 0.9951 - val_loss: 0.3897 - val_accuracy: 0.8676\n",
      "Epoch 90/100\n",
      "1020/1020 [==============================] - 5s 4ms/sample - loss: 0.0281 - accuracy: 0.9990 - val_loss: 0.3719 - val_accuracy: 0.8559\n",
      "Epoch 91/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0328 - accuracy: 0.9961 - val_loss: 0.4103 - val_accuracy: 0.8500\n",
      "Epoch 92/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0293 - accuracy: 0.9980 - val_loss: 0.4043 - val_accuracy: 0.8559\n",
      "Epoch 93/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0333 - accuracy: 0.9951 - val_loss: 0.3899 - val_accuracy: 0.8529\n",
      "Epoch 94/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0283 - accuracy: 0.9961 - val_loss: 0.3979 - val_accuracy: 0.8588\n",
      "Epoch 95/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0229 - accuracy: 0.9990 - val_loss: 0.3871 - val_accuracy: 0.8676\n",
      "Epoch 96/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0275 - accuracy: 1.0000 - val_loss: 0.3754 - val_accuracy: 0.8559\n",
      "Epoch 97/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0241 - accuracy: 0.9971 - val_loss: 0.3908 - val_accuracy: 0.8765\n",
      "Epoch 98/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0268 - accuracy: 0.9980 - val_loss: 0.3858 - val_accuracy: 0.8500\n",
      "Epoch 99/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0289 - accuracy: 0.9971 - val_loss: 0.3936 - val_accuracy: 0.8559\n",
      "Epoch 100/100\n",
      "1020/1020 [==============================] - 4s 4ms/sample - loss: 0.0273 - accuracy: 0.9990 - val_loss: 0.4089 - val_accuracy: 0.8559\n",
      "Results of Phase A\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "def plot(H):\n",
    "    plt.style.use(\"ggplot\")\n",
    "    plt.figure()\n",
    "    plt.plot(np.arange(0, 100), H.history[\"loss\"], label=\"train_loss\")\n",
    "    plt.plot(np.arange(0, 100), H.history[\"val_loss\"], label=\"val_loss\")\n",
    "    plt.plot(np.arange(0, 100), H.history[\"accuracy\"], label=\"train_acc\")\n",
    "    plt.plot(np.arange(0, 100), H.history[\"val_accuracy\"], label=\"val_acc\")\n",
    "    plt.title(\"Training Loss and Accuracy\")\n",
    "    plt.xlabel(\"Epoch #\")\n",
    "    plt.ylabel(\"Loss/Accuracy\")\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "batch_size = 32\n",
    "NUM_EPOCHS = 100\n",
    "\n",
    "vggModel.trainable = False\n",
    "model = tf.keras.models.Sequential()\n",
    "model.add(vggModel)\n",
    "model.add(tf.keras.layers.Flatten())\n",
    "model.add(tf.keras.layers.Dense(256, activation='relu'))  \n",
    "model.add(tf.keras.layers.Dropout(0.2))\n",
    "model.add(tf.keras.layers.Dense(128, activation = 'relu'))\n",
    "model.add(tf.keras.layers.Dropout(0.2))\n",
    "model.add(tf.keras.layers.Dense(17, activation='softmax'))\n",
    "\n",
    "opt = tf.keras.optimizers.SGD(lr=0.01)\n",
    "fname = \"weights.hdf5\"\n",
    "checkpoint1 = tf.keras.callbacks.ModelCheckpoint(fname, monitor=\"val_loss\", mode=\"min\", save_best_only=True, verbose=0)\n",
    "model.compile(loss='sparse_categorical_crossentropy', optimizer=opt, metrics=['accuracy'])\n",
    "H =model.fit(trainX, trainY, epochs=NUM_EPOCHS, batch_size=batch_size, validation_data=(testX, testY), callbacks=[checkpoint1])\n",
    "\n",
    "print(\"Results of Phase A\")\n",
    "plot(H)\n",
    "       \n",
    "                      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:sample_weight modes were coerced from\n",
      "  ...\n",
      "    to  \n",
      "  ['...']\n",
      "Train for 31 steps, validate on 340 samples\n",
      "Epoch 1/100\n",
      "31/31 [==============================] - 9s 289ms/step - loss: 0.4488 - accuracy: 0.8613 - val_loss: 0.3924 - val_accuracy: 0.8588\n",
      "Epoch 2/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.4297 - accuracy: 0.8715 - val_loss: 0.3867 - val_accuracy: 0.8618\n",
      "Epoch 3/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.4294 - accuracy: 0.8725 - val_loss: 0.3832 - val_accuracy: 0.8647\n",
      "Epoch 4/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.4227 - accuracy: 0.8644 - val_loss: 0.3835 - val_accuracy: 0.8647\n",
      "Epoch 5/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.4047 - accuracy: 0.8725 - val_loss: 0.3833 - val_accuracy: 0.8676\n",
      "Epoch 6/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3870 - accuracy: 0.8755 - val_loss: 0.3871 - val_accuracy: 0.8676\n",
      "Epoch 7/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3487 - accuracy: 0.8816 - val_loss: 0.3863 - val_accuracy: 0.8647\n",
      "Epoch 8/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3165 - accuracy: 0.8957 - val_loss: 0.3860 - val_accuracy: 0.8647\n",
      "Epoch 9/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3692 - accuracy: 0.8755 - val_loss: 0.3880 - val_accuracy: 0.8676\n",
      "Epoch 10/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.4068 - accuracy: 0.8735 - val_loss: 0.3892 - val_accuracy: 0.8676\n",
      "Epoch 11/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3429 - accuracy: 0.8866 - val_loss: 0.3875 - val_accuracy: 0.8706\n",
      "Epoch 12/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3376 - accuracy: 0.8988 - val_loss: 0.3886 - val_accuracy: 0.8706\n",
      "Epoch 13/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3362 - accuracy: 0.8897 - val_loss: 0.3904 - val_accuracy: 0.8647\n",
      "Epoch 14/100\n",
      "31/31 [==============================] - 8s 242ms/step - loss: 0.3680 - accuracy: 0.8856 - val_loss: 0.3888 - val_accuracy: 0.8706\n",
      "Epoch 15/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3275 - accuracy: 0.9018 - val_loss: 0.3887 - val_accuracy: 0.8706\n",
      "Epoch 16/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3551 - accuracy: 0.8866 - val_loss: 0.3875 - val_accuracy: 0.8706\n",
      "Epoch 17/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.3519 - accuracy: 0.8866 - val_loss: 0.3882 - val_accuracy: 0.8706\n",
      "Epoch 18/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3117 - accuracy: 0.9008 - val_loss: 0.3857 - val_accuracy: 0.8765\n",
      "Epoch 19/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3116 - accuracy: 0.9049 - val_loss: 0.3860 - val_accuracy: 0.8765\n",
      "Epoch 20/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3205 - accuracy: 0.8917 - val_loss: 0.3865 - val_accuracy: 0.8735\n",
      "Epoch 21/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3565 - accuracy: 0.8947 - val_loss: 0.3844 - val_accuracy: 0.8735\n",
      "Epoch 22/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3753 - accuracy: 0.8755 - val_loss: 0.3826 - val_accuracy: 0.8765\n",
      "Epoch 23/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3370 - accuracy: 0.8911 - val_loss: 0.3822 - val_accuracy: 0.8765\n",
      "Epoch 24/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3600 - accuracy: 0.8877 - val_loss: 0.3806 - val_accuracy: 0.8765\n",
      "Epoch 25/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2974 - accuracy: 0.9008 - val_loss: 0.3832 - val_accuracy: 0.8735\n",
      "Epoch 26/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3296 - accuracy: 0.8897 - val_loss: 0.3853 - val_accuracy: 0.8735\n",
      "Epoch 27/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3381 - accuracy: 0.8877 - val_loss: 0.3848 - val_accuracy: 0.8765\n",
      "Epoch 28/100\n",
      "31/31 [==============================] - 8s 242ms/step - loss: 0.3528 - accuracy: 0.8806 - val_loss: 0.3826 - val_accuracy: 0.8765\n",
      "Epoch 29/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.2925 - accuracy: 0.9049 - val_loss: 0.3832 - val_accuracy: 0.8765\n",
      "Epoch 30/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3453 - accuracy: 0.8978 - val_loss: 0.3824 - val_accuracy: 0.8765\n",
      "Epoch 31/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3271 - accuracy: 0.8927 - val_loss: 0.3810 - val_accuracy: 0.8824\n",
      "Epoch 32/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2941 - accuracy: 0.9099 - val_loss: 0.3788 - val_accuracy: 0.8824\n",
      "Epoch 33/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3091 - accuracy: 0.8897 - val_loss: 0.3788 - val_accuracy: 0.8794\n",
      "Epoch 34/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2968 - accuracy: 0.9099 - val_loss: 0.3778 - val_accuracy: 0.8794\n",
      "Epoch 35/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3108 - accuracy: 0.9069 - val_loss: 0.3772 - val_accuracy: 0.8794\n",
      "Epoch 36/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3241 - accuracy: 0.8917 - val_loss: 0.3792 - val_accuracy: 0.8794\n",
      "Epoch 37/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3505 - accuracy: 0.8887 - val_loss: 0.3773 - val_accuracy: 0.8794\n",
      "Epoch 38/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2988 - accuracy: 0.9049 - val_loss: 0.3781 - val_accuracy: 0.8824\n",
      "Epoch 39/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3080 - accuracy: 0.8927 - val_loss: 0.3763 - val_accuracy: 0.8794\n",
      "Epoch 40/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.3338 - accuracy: 0.8978 - val_loss: 0.3757 - val_accuracy: 0.8853\n",
      "Epoch 41/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3414 - accuracy: 0.8917 - val_loss: 0.3733 - val_accuracy: 0.8853\n",
      "Epoch 42/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2966 - accuracy: 0.8957 - val_loss: 0.3737 - val_accuracy: 0.8853\n",
      "Epoch 43/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3009 - accuracy: 0.8988 - val_loss: 0.3729 - val_accuracy: 0.8853\n",
      "Epoch 44/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3290 - accuracy: 0.8988 - val_loss: 0.3702 - val_accuracy: 0.8824\n",
      "Epoch 45/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2985 - accuracy: 0.8988 - val_loss: 0.3678 - val_accuracy: 0.8853\n",
      "Epoch 46/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3163 - accuracy: 0.9028 - val_loss: 0.3656 - val_accuracy: 0.8853\n",
      "Epoch 47/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3320 - accuracy: 0.9038 - val_loss: 0.3675 - val_accuracy: 0.8882\n",
      "Epoch 48/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2752 - accuracy: 0.9079 - val_loss: 0.3657 - val_accuracy: 0.8853\n",
      "Epoch 49/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2846 - accuracy: 0.9150 - val_loss: 0.3620 - val_accuracy: 0.8853\n",
      "Epoch 50/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2775 - accuracy: 0.9079 - val_loss: 0.3633 - val_accuracy: 0.8853\n",
      "Epoch 51/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2739 - accuracy: 0.9099 - val_loss: 0.3624 - val_accuracy: 0.8853\n",
      "Epoch 52/100\n",
      "31/31 [==============================] - 8s 242ms/step - loss: 0.3185 - accuracy: 0.8877 - val_loss: 0.3626 - val_accuracy: 0.8882\n",
      "Epoch 53/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3013 - accuracy: 0.9038 - val_loss: 0.3613 - val_accuracy: 0.8882\n",
      "Epoch 54/100\n",
      "31/31 [==============================] - 8s 242ms/step - loss: 0.3161 - accuracy: 0.8937 - val_loss: 0.3593 - val_accuracy: 0.8882\n",
      "Epoch 55/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3138 - accuracy: 0.8957 - val_loss: 0.3585 - val_accuracy: 0.8882\n",
      "Epoch 56/100\n",
      "31/31 [==============================] - 8s 242ms/step - loss: 0.2939 - accuracy: 0.9008 - val_loss: 0.3587 - val_accuracy: 0.8882\n",
      "Epoch 57/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2756 - accuracy: 0.9099 - val_loss: 0.3565 - val_accuracy: 0.8882\n",
      "Epoch 58/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2665 - accuracy: 0.9018 - val_loss: 0.3575 - val_accuracy: 0.8882\n",
      "Epoch 59/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2970 - accuracy: 0.9038 - val_loss: 0.3565 - val_accuracy: 0.8882\n",
      "Epoch 60/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3319 - accuracy: 0.8887 - val_loss: 0.3570 - val_accuracy: 0.8882\n",
      "Epoch 61/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2839 - accuracy: 0.9119 - val_loss: 0.3559 - val_accuracy: 0.8882\n",
      "Epoch 62/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.2718 - accuracy: 0.9150 - val_loss: 0.3558 - val_accuracy: 0.8882\n",
      "Epoch 63/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2942 - accuracy: 0.9038 - val_loss: 0.3561 - val_accuracy: 0.8882\n",
      "Epoch 64/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2778 - accuracy: 0.9089 - val_loss: 0.3556 - val_accuracy: 0.8882\n",
      "Epoch 65/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3007 - accuracy: 0.9049 - val_loss: 0.3551 - val_accuracy: 0.8882\n",
      "Epoch 66/100\n",
      "31/31 [==============================] - 7s 242ms/step - loss: 0.2933 - accuracy: 0.8998 - val_loss: 0.3530 - val_accuracy: 0.8882\n",
      "Epoch 67/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.3319 - accuracy: 0.8907 - val_loss: 0.3545 - val_accuracy: 0.8882\n",
      "Epoch 68/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2822 - accuracy: 0.9079 - val_loss: 0.3559 - val_accuracy: 0.8882\n",
      "Epoch 69/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.3042 - accuracy: 0.8937 - val_loss: 0.3559 - val_accuracy: 0.8882\n",
      "Epoch 70/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.3048 - accuracy: 0.9119 - val_loss: 0.3553 - val_accuracy: 0.8882\n",
      "Epoch 71/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2846 - accuracy: 0.9133 - val_loss: 0.3554 - val_accuracy: 0.8882\n",
      "Epoch 72/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2921 - accuracy: 0.8998 - val_loss: 0.3541 - val_accuracy: 0.8882\n",
      "Epoch 73/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2792 - accuracy: 0.9049 - val_loss: 0.3531 - val_accuracy: 0.8882\n",
      "Epoch 74/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.3145 - accuracy: 0.8917 - val_loss: 0.3499 - val_accuracy: 0.8882\n",
      "Epoch 75/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2410 - accuracy: 0.9251 - val_loss: 0.3517 - val_accuracy: 0.8882\n",
      "Epoch 76/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.3082 - accuracy: 0.8927 - val_loss: 0.3494 - val_accuracy: 0.8882\n",
      "Epoch 77/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2596 - accuracy: 0.9211 - val_loss: 0.3504 - val_accuracy: 0.8912\n",
      "Epoch 78/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2841 - accuracy: 0.9194 - val_loss: 0.3490 - val_accuracy: 0.8882\n",
      "Epoch 79/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.3004 - accuracy: 0.8988 - val_loss: 0.3504 - val_accuracy: 0.8882\n",
      "Epoch 80/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2585 - accuracy: 0.9130 - val_loss: 0.3522 - val_accuracy: 0.8824\n",
      "Epoch 81/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2750 - accuracy: 0.9180 - val_loss: 0.3515 - val_accuracy: 0.8853\n",
      "Epoch 82/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.3152 - accuracy: 0.8957 - val_loss: 0.3487 - val_accuracy: 0.8882\n",
      "Epoch 83/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2768 - accuracy: 0.9140 - val_loss: 0.3466 - val_accuracy: 0.8912\n",
      "Epoch 84/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2572 - accuracy: 0.9150 - val_loss: 0.3460 - val_accuracy: 0.8853\n",
      "Epoch 85/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2653 - accuracy: 0.9190 - val_loss: 0.3477 - val_accuracy: 0.8912\n",
      "Epoch 86/100\n",
      "31/31 [==============================] - 8s 247ms/step - loss: 0.2507 - accuracy: 0.9170 - val_loss: 0.3454 - val_accuracy: 0.8882\n",
      "Epoch 87/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2684 - accuracy: 0.9160 - val_loss: 0.3445 - val_accuracy: 0.8912\n",
      "Epoch 88/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.2876 - accuracy: 0.9059 - val_loss: 0.3444 - val_accuracy: 0.8882\n",
      "Epoch 89/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2657 - accuracy: 0.9200 - val_loss: 0.3441 - val_accuracy: 0.8882\n",
      "Epoch 90/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2405 - accuracy: 0.9302 - val_loss: 0.3451 - val_accuracy: 0.8853\n",
      "Epoch 91/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2305 - accuracy: 0.9271 - val_loss: 0.3457 - val_accuracy: 0.8853\n",
      "Epoch 92/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.2854 - accuracy: 0.9140 - val_loss: 0.3456 - val_accuracy: 0.8853\n",
      "Epoch 93/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2611 - accuracy: 0.9130 - val_loss: 0.3437 - val_accuracy: 0.8853\n",
      "Epoch 94/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2547 - accuracy: 0.9221 - val_loss: 0.3445 - val_accuracy: 0.8853\n",
      "Epoch 95/100\n",
      "31/31 [==============================] - 8s 247ms/step - loss: 0.2735 - accuracy: 0.9109 - val_loss: 0.3432 - val_accuracy: 0.8853\n",
      "Epoch 96/100\n",
      "31/31 [==============================] - 8s 244ms/step - loss: 0.2872 - accuracy: 0.9079 - val_loss: 0.3427 - val_accuracy: 0.8853\n",
      "Epoch 97/100\n",
      "31/31 [==============================] - 8s 243ms/step - loss: 0.2692 - accuracy: 0.9059 - val_loss: 0.3419 - val_accuracy: 0.8853\n",
      "Epoch 98/100\n",
      "31/31 [==============================] - 8s 245ms/step - loss: 0.2819 - accuracy: 0.9089 - val_loss: 0.3385 - val_accuracy: 0.8882\n",
      "Epoch 99/100\n",
      "31/31 [==============================] - 8s 246ms/step - loss: 0.2660 - accuracy: 0.9150 - val_loss: 0.3379 - val_accuracy: 0.8912\n",
      "Epoch 100/100\n",
      "31/31 [==============================] - 8s 247ms/step - loss: 0.2955 - accuracy: 0.9008 - val_loss: 0.3346 - val_accuracy: 0.8882\n",
      "Results of Phase B of fine tuning\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trainDataGenerator = tf.keras.preprocessing.image.ImageDataGenerator(\n",
    "      shear_range=0.2,\n",
    "      zoom_range=0.2,\n",
    "      rotation_range=30,\n",
    "      horizontal_flip=True)\n",
    "train_generator = trainDataGenerator.flow(trainX, trainY, batch_size=batch_size)\n",
    "\n",
    "\n",
    "opt = tf.keras.optimizers.SGD(lr=1e-5)\n",
    "vggModel.trainable = True\n",
    "trainableFlag = False\n",
    "# print(len(vggModel.layers))\n",
    "i=0\n",
    "for layer in vggModel.layers:\n",
    "    if i >= 10:\n",
    "        trainableFlag = True\n",
    "    layer.trainable = trainableFlag\n",
    "    i+=1\n",
    "    \n",
    "model.compile(loss='sparse_categorical_crossentropy',optimizer=opt, metrics=['accuracy'])\n",
    "\n",
    "\n",
    "H =model.fit(train_generator, epochs=NUM_EPOCHS, validation_data=(testX, testY), \n",
    "            steps_per_epoch=len(trainX)// batch_size, verbose =1 )\n",
    "\n",
    "print(\"Results of Phase B of fine tuning\")\n",
    "\n",
    "plot(H)\n",
    "\n",
    "                 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}