Pretrained_vgg_16.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "OnijNloZyn0d"
   },
   "outputs": [],
   "source": [
    "import tensorflow\n",
    "# from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "XKMYuS6Qys-7",
    "outputId": "dc539377-1cc5-4597-fc9e-1e36d6f50c1f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mounted at /content/drive\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "fy4cM_k0yn0n"
   },
   "outputs": [],
   "source": [
    "import keras\n",
    "from tensorflow.keras.datasets import cifar10\n",
    "from tensorflow.keras import models,layers\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Conv2D, MaxPool2D, Flatten\n",
    "from tensorflow.keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "DSI0lsbfyn0o"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from __future__ import print_function\n",
    "from keras.preprocessing.image import ImageDataGenerator, load_img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "BqUm1itYyn0o"
   },
   "outputs": [],
   "source": [
    "train_dir = '/content/drive/MyDrive/CNN/Dataset/train_images'\n",
    "validation_dir = '/content/drive/MyDrive/CNN/Dataset/test_images'\n",
    "image_size = 224"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "d4c483BOyn0o"
   },
   "source": [
    "# Train last 4 layers with data augmentation\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "r4qQQDk9yn0p"
   },
   "source": [
    "## Build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "9RRH7XtWyn0p",
    "outputId": "f284bc8b-b419-442d-d725-2e888f843a33"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_5\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "vgg16 (Functional)           (None, 7, 7, 512)         14714688  \n",
      "_________________________________________________________________\n",
      "flatten_5 (Flatten)          (None, 25088)             0         \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 1024)              25691136  \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 10)                10250     \n",
      "=================================================================\n",
      "Total params: 40,416,074\n",
      "Trainable params: 32,780,810\n",
      "Non-trainable params: 7,635,264\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.applications import VGG16\n",
    "\n",
    "#Load the VGG model\n",
    "vgg_conv = VGG16(weights='imagenet', include_top=False, input_shape=(image_size, image_size, 3))\n",
    "\n",
    "# Freeze all the layers\n",
    "for layer in vgg_conv.layers[:-4]:\n",
    "    layer.trainable = False\n",
    "\n",
    "# Check the trainable status of the individual layers\n",
    "# for layer in vgg_conv.layers:\n",
    "#     print(layer, layer.trainable)\n",
    "\n",
    "\n",
    "from keras import models\n",
    "from keras import layers\n",
    "from keras import optimizers\n",
    "\n",
    "# Create the model\n",
    "model = models.Sequential()\n",
    "\n",
    "# Add the vgg convolutional base model\n",
    "model.add(vgg_conv)\n",
    "\n",
    "# Add new layers\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dense(1024, activation='relu'))\n",
    "model.add(layers.Dropout(0.5))\n",
    "model.add(layers.Dense(10, activation='softmax'))\n",
    "\n",
    "#Show a summary of the model. Check the number of trainable parameters\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "b8Ml-WPcyn0q"
   },
   "source": [
    "## Training step\n",
    "Train the model. Here we will be using the imageDataGenerator for data augmentation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OPDEoY4yyn0r",
    "outputId": "7d34fa40-9c93-4c79-f497-0d6d955b44f8",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2408 images belonging to 10 classes.\n",
      "Found 604 images belonging to 10 classes.\n",
      "Epoch 1/50\n",
      "76/76 [==============================] - 281s 4s/step - loss: 3.0628 - accuracy: 0.3350 - val_loss: 1.2362 - val_accuracy: 0.6705\n",
      "Epoch 2/50\n",
      "76/76 [==============================] - 170s 2s/step - loss: 1.2312 - accuracy: 0.6380 - val_loss: 1.4036 - val_accuracy: 0.6209\n",
      "Epoch 3/50\n",
      "76/76 [==============================] - 168s 2s/step - loss: 0.9822 - accuracy: 0.6987 - val_loss: 0.5644 - val_accuracy: 0.8146\n",
      "Epoch 4/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.7622 - accuracy: 0.7584 - val_loss: 0.5710 - val_accuracy: 0.8146\n",
      "Epoch 5/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.6200 - accuracy: 0.8068 - val_loss: 0.4639 - val_accuracy: 0.8560\n",
      "Epoch 6/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.5259 - accuracy: 0.8338 - val_loss: 0.5823 - val_accuracy: 0.8063\n",
      "Epoch 7/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.5370 - accuracy: 0.8341 - val_loss: 0.5755 - val_accuracy: 0.7831\n",
      "Epoch 8/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.5028 - accuracy: 0.8506 - val_loss: 0.5328 - val_accuracy: 0.8394\n",
      "Epoch 9/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.4294 - accuracy: 0.8745 - val_loss: 0.3731 - val_accuracy: 0.8841\n",
      "Epoch 10/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.4748 - accuracy: 0.8492 - val_loss: 0.4473 - val_accuracy: 0.8742\n",
      "Epoch 11/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.4343 - accuracy: 0.8731 - val_loss: 0.5047 - val_accuracy: 0.8725\n",
      "Epoch 12/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3896 - accuracy: 0.8693 - val_loss: 0.8420 - val_accuracy: 0.7517\n",
      "Epoch 13/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3800 - accuracy: 0.8825 - val_loss: 0.3345 - val_accuracy: 0.8990\n",
      "Epoch 14/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3273 - accuracy: 0.9040 - val_loss: 0.3397 - val_accuracy: 0.9040\n",
      "Epoch 15/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3254 - accuracy: 0.8962 - val_loss: 0.2744 - val_accuracy: 0.9222\n",
      "Epoch 16/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.3437 - accuracy: 0.8914 - val_loss: 0.2438 - val_accuracy: 0.9205\n",
      "Epoch 17/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3619 - accuracy: 0.8951 - val_loss: 0.6573 - val_accuracy: 0.7881\n",
      "Epoch 18/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3030 - accuracy: 0.9027 - val_loss: 0.2445 - val_accuracy: 0.9156\n",
      "Epoch 19/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2836 - accuracy: 0.9153 - val_loss: 0.3248 - val_accuracy: 0.9172\n",
      "Epoch 20/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.3034 - accuracy: 0.9101 - val_loss: 1.1422 - val_accuracy: 0.6772\n",
      "Epoch 21/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2682 - accuracy: 0.9176 - val_loss: 0.3038 - val_accuracy: 0.9073\n",
      "Epoch 22/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.2753 - accuracy: 0.9103 - val_loss: 0.3140 - val_accuracy: 0.9056\n",
      "Epoch 23/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.2473 - accuracy: 0.9284 - val_loss: 0.5847 - val_accuracy: 0.8129\n",
      "Epoch 24/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2759 - accuracy: 0.9084 - val_loss: 0.2240 - val_accuracy: 0.9437\n",
      "Epoch 25/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2742 - accuracy: 0.9195 - val_loss: 0.6467 - val_accuracy: 0.8179\n",
      "Epoch 26/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2624 - accuracy: 0.9210 - val_loss: 0.3793 - val_accuracy: 0.9040\n",
      "Epoch 27/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2759 - accuracy: 0.9239 - val_loss: 0.3650 - val_accuracy: 0.8891\n",
      "Epoch 28/50\n",
      "76/76 [==============================] - 165s 2s/step - loss: 0.2605 - accuracy: 0.9190 - val_loss: 0.3187 - val_accuracy: 0.9172\n",
      "Epoch 29/50\n",
      "76/76 [==============================] - 165s 2s/step - loss: 0.2454 - accuracy: 0.9270 - val_loss: 0.2758 - val_accuracy: 0.9238\n",
      "Epoch 30/50\n",
      "76/76 [==============================] - 165s 2s/step - loss: 0.2374 - accuracy: 0.9284 - val_loss: 0.3347 - val_accuracy: 0.8924\n",
      "Epoch 31/50\n",
      "76/76 [==============================] - 165s 2s/step - loss: 0.2694 - accuracy: 0.9181 - val_loss: 0.4154 - val_accuracy: 0.8990\n",
      "Epoch 32/50\n",
      "76/76 [==============================] - 165s 2s/step - loss: 0.2117 - accuracy: 0.9377 - val_loss: 0.8511 - val_accuracy: 0.8113\n",
      "Epoch 33/50\n",
      "76/76 [==============================] - 165s 2s/step - loss: 0.3021 - accuracy: 0.9178 - val_loss: 0.4430 - val_accuracy: 0.8940\n",
      "Epoch 34/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2465 - accuracy: 0.9281 - val_loss: 0.3760 - val_accuracy: 0.9123\n",
      "Epoch 35/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2217 - accuracy: 0.9366 - val_loss: 0.3508 - val_accuracy: 0.9040\n",
      "Epoch 36/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2365 - accuracy: 0.9329 - val_loss: 0.3594 - val_accuracy: 0.9040\n",
      "Epoch 37/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.2146 - accuracy: 0.9378 - val_loss: 0.4068 - val_accuracy: 0.8891\n",
      "Epoch 38/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2745 - accuracy: 0.9242 - val_loss: 0.2884 - val_accuracy: 0.9321\n",
      "Epoch 39/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2744 - accuracy: 0.9262 - val_loss: 0.2827 - val_accuracy: 0.9089\n",
      "Epoch 40/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.1993 - accuracy: 0.9435 - val_loss: 0.2486 - val_accuracy: 0.9354\n",
      "Epoch 41/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2747 - accuracy: 0.9159 - val_loss: 0.3012 - val_accuracy: 0.9338\n",
      "Epoch 42/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2319 - accuracy: 0.9402 - val_loss: 0.2575 - val_accuracy: 0.9404\n",
      "Epoch 43/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.2026 - accuracy: 0.9371 - val_loss: 0.3371 - val_accuracy: 0.9288\n",
      "Epoch 44/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2495 - accuracy: 0.9319 - val_loss: 0.2934 - val_accuracy: 0.9288\n",
      "Epoch 45/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2164 - accuracy: 0.9302 - val_loss: 0.2586 - val_accuracy: 0.9338\n",
      "Epoch 46/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.1632 - accuracy: 0.9490 - val_loss: 0.3565 - val_accuracy: 0.9073\n",
      "Epoch 47/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.2192 - accuracy: 0.9368 - val_loss: 0.3005 - val_accuracy: 0.9255\n",
      "Epoch 48/50\n",
      "76/76 [==============================] - 167s 2s/step - loss: 0.2044 - accuracy: 0.9439 - val_loss: 0.2919 - val_accuracy: 0.9387\n",
      "Epoch 49/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.1767 - accuracy: 0.9456 - val_loss: 0.3907 - val_accuracy: 0.8940\n",
      "Epoch 50/50\n",
      "76/76 [==============================] - 166s 2s/step - loss: 0.2086 - accuracy: 0.9428 - val_loss: 0.4548 - val_accuracy: 0.9089\n"
     ]
    }
   ],
   "source": [
    "train_datagen = ImageDataGenerator(\n",
    "      rescale=1./255,\n",
    "      rotation_range=20,\n",
    "      width_shift_range=0.2,\n",
    "      height_shift_range=0.2,\n",
    "      horizontal_flip=True,\n",
    "      fill_mode='nearest'\n",
    "      )\n",
    "\n",
    "\n",
    "validation_datagen = ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "# Change the batchsize according to your system RAM\n",
    "train_batchsize = 32\n",
    "val_batchsize = 32\n",
    "\n",
    "#parameters\n",
    "eps = 1#40\n",
    "\n",
    "# Data Generator for Training data\n",
    "train_generator = train_datagen.flow_from_directory(\n",
    "        train_dir,\n",
    "        target_size=(image_size, image_size),\n",
    "        batch_size=train_batchsize,\n",
    "        class_mode='categorical')\n",
    "\n",
    "# Data Generator for Validation data\n",
    "validation_generator = validation_datagen.flow_from_directory(\n",
    "        validation_dir,\n",
    "        target_size=(image_size, image_size),\n",
    "        batch_size=val_batchsize,\n",
    "        class_mode='categorical',\n",
    "        shuffle=False)\n",
    "\n",
    "cnn = models.Sequential([\n",
    "     layers.Conv2D(filters =32, kernel_size=(3,3),activation = 'relu',input_shape = (224,224,3)),\n",
    "     layers.MaxPooling2D((2,2)),\n",
    "     layers.Conv2D(filters =64, kernel_size=(3,3),activation = 'relu'),\n",
    "     layers.MaxPooling2D((2,2)),\n",
    "     # dense layer\n",
    "     layers.Flatten(),\n",
    "     layers.Dense(64,activation = 'relu'),\n",
    "     layers.Dense(10,activation = 'softmax')\n",
    "     ])\n",
    "# Compile the model\n",
    "cnn.compile(loss='categorical_crossentropy',\n",
    "              optimizer=optimizers.RMSprop(lr=0.001),\n",
    "              metrics=['accuracy'])\n",
    "# cnn.compile(optimizer ='adam',\n",
    "#             loss = 'sparse_categorical_crossentropy',\n",
    "#             metrics=['accuracy'])\n",
    "history = cnn.fit(train_generator,\n",
    "                              epochs=50,\n",
    "                              verbose=1,\n",
    "                              validation_data=validation_generator)\n",
    "\n",
    "model.save('trained_last4_layers_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "U-FWBn1hyn0r"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 590
    },
    "id": "uV3hWpXJyn0s",
    "outputId": "a2e6c347-da7e-4744-c228-d5c0067114ff"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2dd5hU5fXHP4elC9IV6SBWkKKAYold0BgRTLB3xcQSe409mpho7CbGQvSnYkNBo0ixYgsIAgoKArrIIsrSXXCB3T2/P85cdpidmZ2ZndnZ3Tmf55ln5t77zr3n3p19v+855y2iqjiO4zi5S71sG+A4juNkFxcCx3GcHMeFwHEcJ8dxIXAcx8lxXAgcx3FyHBcCx3GcHMeFwEkaEXlLRM5Md9lsIiL5InJElq5dK56RU3cRH0eQG4hIUdhmU2ATUBravkBVn6t+q2oOIpIPnKeqb0fsfws4KLTZCFBgc2j7WVX9fZLXuRXoqaqnVclgx0kj9bNtgFM9qGqz4HOsSi90rL6qllSnbTUZVT06+CwiTwEFqnpj9iyqPfhvqfbgoaEcR0QOEZECEblWRH4E/iMirUTkDREpFJE1oc+dwr7zvoicF/p8loh8JCL3hMp+JyJHp1i2u4hMFZGfReRtEXlERJ6NYXciNv5ZRD4OnW+yiLQNO366iCwRkVUi8qcUn92xIjJbRNaKyCci0ifs2LUisix07QUicriIDAVuAE4UkSIRmZPlZ9RaRP4jIj+Ejo8POzYsdG/rRWRxyPYKITQRuTW4voh0ExEVkXNF5Hvg3dD+l0XkRxFZF7K9V9j3m4jIP0J/i3Whe28iIm+KyCUR9/OFiAxP5W/lxMeFwAFoD7QGugKjsN/Ff0LbXYBfgIfjfH9fYAHQFvg78KSISAplxwDTgTbArcDpca6ZiI2nAGcDOwANgasARGRP4F+h83cIXa8TSSAi/YHRwAWh7/8beF1EGonIbsDFwEBVbQ4MAfJVdSLwF+BFVW2mqn1jnL66ntEzWJiwF/aM7gvd2yDg/4CrgZbAr4D8Sh9KOQcDe2D3DfAWsEvoGp8D4WHIe4B9gP2x3+A1QBnwNLA1fCYifYGOwJtJ2OEkiqr6K8de2D/1EaHPh2Ax78ZxyvcD1oRtv4+FlgDOAhaFHWuKxdHbJ1MWq6hKgKZhx5/F4vCJ3FM0G28M274QmBj6fDPwQtix7ULP4IhKrvEUcEfo87+AP0ccX4BVgj2BFcARQIOIMrdG3lM2nhGwE1bhtopS7t/AfZX9diLvB+gWsrVHHBtahsq0wITqF6BvlHKNgTXALqHte4B/Zvt/p66+3CNwAApVtTjYEJGmIvLvkLu+HpgKtBSRvBjf/zH4oKobQx+bJVm2A7A6bB/A0lgGJ2jjj2GfN4bZ1CH83Kq6AVgV61ox6ApcGQoLrRWRtUBnoIOqLgIuwyrJFSLygoh0SOLc1fGMOofOtSbKVzsDi5OwN5KtNolInojcFQovrafcs2gbejWOdq3Q7/FF4DQRqQecjHkwTgZwIXDAWmjhXAnsBuyrqttjoQGAWOGedLAcaC0iTcP2dY5Tvio2Lg8/d+iabZIzl6XAnaraMuzVVFWfB1DVMap6ICYYCvwt9L2qdNNL5zNaGjpXyyjfWwrsHOOcGzAvJaB9lDLh93gKMAzzjlpgXkNgw0qgOM61ngZOBQ4HNqrqpzHKOVXEhcCJRnPMZV8rIq2BWzJ9QVVdAswAbhWRhiIyGPhNhmwcCxwrIgeKSEPgdpL/X3gc+L2I7CvGdiLyaxFpLiK7ichhItIIq+h+wcIwAD8B3UKt3KRI5zNS1eVY7P6foaRyAxEJhOJJ4OxQgrueiHQUkd1Dx2YDJ4XKDwB+W4nZzbGuyqswAflLmA1lWJ7lXhHpEPIeBoeeG6GKvwz4B+4NZBQXAica9wNNsBbb/4CJ1XTdU4HBWKVxBxYa2BSjbMo2quo84CIs8boci0UXJGOoqs4AzseSr2uARVh8H2y8wV0h237EkqTXh469HHpfJSKfJ3PNEOl8RqcDW4D5WE7jstC9TceS7PcB64APMM8G4CasBb8GuA17hvH4P2AJsAz4KmRHOFcBXwKfAasxz6lexPf3wnIhTobwAWVOjUVEXgTmq2rGPZLaSl1/RiJyBjAqFGZzMoR7BE6NQUQGisjOoXDEUCy2PL6y7+USufSMQrmQC4HHsm1LXcdHFjs1ifbAq1jitgD4g6rOyq5JNY6ceEYiMgS7z7epPPzkVBEPDTmO4+Q4HhpyHMfJcWpdaKht27barVu3bJvhOI5Tq5g5c+ZKVW0X7VitE4Ju3boxY8aMbJvhOI5TqxCRJbGOeWjIcRwnx3EhcBzHyXFcCBzHcXIcFwLHcZwcx4XAcRwnx3EhcBzHyXFcCBzHcXIcFwLHqYSFC2HSpGxb4TiZw4XAcSrhzjth5EjwablqJ6tXw8aNlZfLZVwIHKcSvvsO1q+HNdFW93VqJKowdSqceCLsuCP06AFTpmToYps3w3PPwSuvmOpURlkZs8Yu5pnr5lVevrQUpk2DW26BgQPh9dfTY3MEtW6KCcepbpaEBubn50Pr1lk1pfaxZQt8+CG88Qa0bw/nnVflhzhzJjRpAh06QIsWIGGrVBcVwbPPwiOPwNy50LIlXHQRvP02HHUUXHst/PnP0KBBFe8LrJJ+7jmrpPPzbZ8I9O8Phx9ur4MOgvr1zeiPPoIPP+SFd3fg7A0PUUwTBv+tJz13BgYNsop+0CDo1g0++ADeegsmToSVK6FePdhvPztXJlDVWvXaZ5991HGqiy1bVPPyVEF17NhsW1MFfv5ZdeRI1YEDVY8/XvWii1T/8hfVp55SnTJFtaAgfdfasEH11VdVzzhDtVUre3iNGtl706Z27W++Sf68ZWX64g2z1Nr75afbZRfVQw5RHTFCdfvtbX///qpPPGGmBCaNGmXH9t1X9dvP16i++669SksrvfTGjaqvv6569dWqBUvL7P723LP8YhMmqH78sertt6sefLBqgwZ2rEED1caNVUFLEb259UMKqoN6rFCRMr3tiKlmeKdOus2NgWrbtqqnn676/POqq1Yl/7wiAGZojHo16xV7si8XAqc6WbKk/P/ynnuybU2KFBWp/upXpmiHH67aq5dqy5YVK55dd1X9wx9M8ZKteIqLVV95RXX4cNUmTex8rVqZGIwbZzXxF1+onnOOasOGqiKqw4bp7ed8p4cdVqabNsU5d0mJ6ksv6c99D9COLNX+zNTn216s99y0Vi+/XPXEE1UPPFC1Z0/VU09V/eQT1bKysO//9JPqiy+q3nCDvrT3X7SFrNXtWasvMNLs3H131X//22r7MFavVv2//7N6umnT8sc0rOV79mG33VRffjm6kBQVqU6cqHrNNaqXX64bxozXkcf9oqB61ln2uA4+2B75Vlt/+EH1tddU771Xddq0hAQqGVwIHCdFpk4trwAuvjjb1qTAxo1W+derpzpmzLbHNmywlvl771nlc8wxqs2a2c2KqO69t+oVV6g+84xV4ps3b/v9sjLVTz818Qha/jvtpHrJJarvvFOxfMDy5ao33aSbWrfX1qxUUL1xz1fMQ3nzTfNOysqstnz8cWvyg17T6t8Kqp/8e47ZueeeqitXxr//999Xbd3abMvLU+3dW78bdqnu12WZguqRvX/QkS0n6Uhe0JENx+nIPb/Ukcf9ooccopqXV6ag2qHFz3ph1zd0cpPj9A5uUFCdcNkkcxcTYNky1QED7JH+/e/lFf9jj5lZn32W0GmqjAuB46TIM8/Yf8l226kee2wWDMjPt9ZhRGs1IX75RXXIEKuBnn46se9s3qz60Ueqt91mXkTDhuVK2LChar9+qmeeaS3dUAWtTZqonnKKtYBLShI2b9wLxQqqe23/neaxRacxsPxabdqo7rCDBqGXr+97S+vXL9Ozzw59+d13Ldw0aJCFvaLx5JMWmtl9d3MTiou3uc0bb1TdYw/V3Xcv0927FOnuzb7X3flKd+dr3Xv7b/T6pvfrNAZqKWL3+vvf66YxY3W3XUu1Z89tTheTmTNVO3Sw389rr217bPVqe6SXX57wI6sSLgROxonr2tdi7rjD/kuOPFK1d+8MX6y01Fre//yn6skn648d+usd3KBX8zcty6tvsehRoyz4/cUX8SvdTZtUf/1rM/6JJ1K3afNm1blzVZ97zir/IUNU27e38x5yiOro0arr1qV06uHDVXfc0Rr1nTur7rZLiW6c8pHqQw+pnnee6m9/qzppkpaVlukRR1g066efwk4wfnx5uCu8Vi4pUb3qKrPxqKNU16xJ3KgFC1QvvNBiNqeeave3ZMk2RSZPtlPfcUf8U82fb85I586qc+ZEL3P88eZEJaGfKeNC4GSMsjJr0YAl6vbYQ/WII6zReP319n9UHT/yTHH++art2qleeqm16srK1CqWKVMslHHBBaoPP6w6fXpiTcRwiout9X3XXVZpt2ypZaAfM1hPafyKNqi3ZWsD+c2RT9mDbdGivNW83XZWCd56q+rbb1tcWtUq7+OPtzL/+lfan4mqmrdRBVautMb6FVfY9ttvm7mXXlqx7Msv27GHHopyoqeftoMjRlioZv161d/8RrfG8hIM3yTLCSeYI5SfH/34ihWqPXrYb2fRotjneeklM3XKlIyYuQ0uBE7GuOce+xWNHGmh4REjrFdG586q9evbsY8+qto1yspU166tRq+jrMwqlG+/1aP2XasDdlmj9w3/QEG1cOd9yyti2LZibtjQeuVcdJEp4EsvVXw9/7wp5EEHlfekAd2waz99/MCntF+XVSFRLdM//tEa/t26WYy5rEzNa1iwwGJWF12k2revhX6CGPjAgZY5BdUHH6ymB5Y8Dz9sJs6eXb7v4ott37vvlu8rKrLfUt++cer0+++3L558supee9lzeOSRjNq/ZIklkEeMqHhs40bV/fazzkKffhr/PBs3qjZvruUhrwziQuBkhKCl9rvfRe/gMH++HX/mmcTPOWWKRSBOPdUiD7vuag1fsPotYd580/zyc85RXbw4ftmyMksq/vrXFtANi4vvxtd6Ai/reI6zxN6vrrCYwOTJFuQtK1P9/nvraXPNNWZ0kHCN9apf32LbV16pOm6clv20Ymvdvddeqo8+um3Y+/HH7dh//xvD/rVrrfviDTdYXL9tW9X77kviYVU/gwap9umz7b4NGywU37VrebTp+usTbEzccku5ME+enAGLK3LnnXbJSZPK95WUmDiIWCeqRDjrLPOmq+hkVYoLgZN2Pv7YGrT77x/7B7xxoyYUSw2nY0cLGXTrpnrAAeZpXH65VRw77pjgST75xPz2rl3NyLy86IJQWmqdwwcPNkN32MGaZtdeq3r33Vo2+j/apGGJXnHyMp096UcFE79KKSmx3jhz50Z/BSGcEM8/r1tDH9t0ewyxebNq9+6q++wT/XimKC21EEe6+fpru99//KPisU8+sQ5O555rjk+DBhZmrJSyMusVlcr4hBQpLjbh2nXX8qjgFVfYvd17b+LnCXIOCf22qoALgZNWFi60RmfPnqqFhfHLtmtn+c1EKC62X+Rtt1U8du211lCvtCKcN8+6MvbsaZnFZctU//jHckE4+2xzVZ55xrK/YKrzyCMVeub89JMdfuABSwuA6t13J3YvibJxo2qXLpYHjtdt/Mkn7fqvv57e68eitNQ6AtWrZxXxd9+l79w33GDnXb48+vHrrtOtwxpatFD98cf0XTvdvPWW2frXv5qQg4VIkxHskhLLvw8fHr9cKh3HwnEhcNJGYaHVsW3amCBUxoAB1tEkEb75xn6R0Xo6/vXOEoXykaJR+f57G6HZvr3qt9/q/PlWgW3apCYIl166dZSngg2seuaZmP3dp0+3YuPH23bLltahJJ389a9aIS4ejc2bLfm4997V4xUEIZkhQ0xDGzSwlMQPP1TtvKWlFvM/+ujYZYqLLWwUiHBN5/jj7WdVr57qsGGpdY647DJr6KxeHf34/Pnm4L70Uup2Zk0IgKHAAmARcF2U412Bd4AvgPeBTpWd04Uge/zyi4VrGjWy0FAijBhh3bgTIXCRP/ggtKOkxAYmnX++PrrdFQqqBZfcFdGHMMTKldZlafvtVWfP1jlzyut7MM+kb1/VYw7/RS8YNFMXPvp2pSM3gxzIrFm23b+/jblKFz/+aInCYcMSK/+f/2wrTJkiGOg0apSJztKl1jmqfn2LuF17beozHrz7rp37+efjl/vmG4vBZ6jTT1r57jt7LgMHVoj6JUzQ6Hj88YrHZs2y3+8OO2ybXE+WrAgBkAcsBnoADYE5wJ4RZV4Gzgx9Pgx4prLzuhBUL2Vlql99ZT0lg1ZaMq2Syy+33hWJtGKDCmjJy9Os+bnjjhp0k3zxgPsVVL+ktzW/Lr64vO9eUZF102jUyJK+aiEUUP3TnyzUNGqUDQjr399abkG3xXgEPaKCVtrw4TaYtTJWrrTcyQsvxC8XVK4LFlR+TlWrFHv2tDFdmfIK3nrLImhDh1ashBcutCS+iD3qtm0rvtq3t048sTjzTNPqqoY5ahr5+ZV4q5VQVmb5hkMP3Xb/xx9beKxz58R/J7HIlhAMBiaFbV8PXB9RZh7QOfRZgPWVndeFIPOUllq3t2uusTht0KoeNMjGFSVD0LOvslyCfv21/mn/dzWPLbqFPKvsTzjBVGfDhq3ewtRn8i3x26CB1Vinn14+ejasm0bQvTxa+Kp//8TCVZdcYi32oNJNVNTGj7dri9gUNtH48ksTpGj95uMR3Nerryb3vUSYPds6PPXta71nY/HllyakF15Y8XXooWbfzTdXfE5FRdYD7Lzz0m97XeDWW+03E8z/N3ly+aR6EWPaUiJbQvBb4Imw7dOBhyPKjAEuDX0eASjQJsq5RgEzgBldunSp+hNx4nLAAbq1l+ORR1oeNdXJKV991c41c2aUg4WFlmEbaFMLnCrParcmy1WffbZCTfTZZ3aercP0v//eAqvBbGCPPrpN+fvus93RQhinn269kyrjuOO2HU38wAN2zmiRqXBuu83+oY86SmMmmIcMsZx2siGWLVusYujbt2Jka8ECE5YuXcLCawmydKk9k06dqjYR6ZYtlo8PBoeF2xhM1zF1aurnr8sEObK777Z5+ho2NC88XcnymiwEHYBXgVnAA0AB0DLeed0jyCxbtlhL9fTTExyZX1ZmHanvvtualBHNwJkzI1qwRUUWIB42rHzEWb9+qvfeqwcM3KSHHBL9MosWWdGnnoo4UFhofQ4juPlmq4yjJe7+9jc7V2X317evDS0IeO01+960afG/N2KEVdabNln3V7B5bYJHE/Q0SbWrf1ChvvKK3d/48SbYYI5SvXrW8yZR1q2zCqd589hTISRDaamJAJgoBH+DI4+0brBpnlSzTjFwoOUC8vIs2hkreZwKNTY0FFG+GVBQ2XldCDLL0qX2q4gV0tjKL79Yn8ZevXSbrGz37hZDmTpVtaRECwtDld7Zs23kWTBF8U472YCqsJqnY0cbXBONVauSqzwvvth6+UTjjTc0oUFKkb2EvvjCvvfii/G/t/PONk2OqlWC552nW7sVbt5seYaePVMfKb1li4XsunSxF9izu/1265LZp09ySe3zz7eKJ3xgVFUpKysf4/Xb36p++60J8803p+8adZEglHrYYbHn0kuVbAlBfeBboHtYsrhXRJm2QL3Q5zuB2ys7rwtBZvn0U/tVvPFGjAIrVliNE8wM2aePNdO//966PBxzTPnI3HbttOzoY7QpRXo5/7CuD3/4g8UtIpqFmzZZRXHrrdEvW1Jip7zllsTu45RTrLtlNPLzNVo0aRvWrbMyf/tb/H2RrF9vZf785/J9ZWXlA4322svex41L7D5iEcxRc+ihNqg5PLF72mmJhb4CevasvA97qtx779afQsycjVNOcbFFRjMxyjieEGRsqUpVLRGRi4FJWA+i0ao6T0RuDxn0OnAI8FcRUWAqcFGm7HESoKyMZWOnAYPpdPqh0HQhNGpU/mrYEL74AoqL4Zhj4Ior4LDDytcKPO88e/38sy2zN348Mn06XVuuY0nv0+C9P8Zcam/pUnMpunaNblpeni1LmOi6wWvWxF4RsUsXaNbMljKMRbA8Zbg9229v5wxWJYzGl1/ae9++5ftE4J57zP5bboGDD4ZhwxK6jZj87ne2jnLz5hWP9e1ryzWuWgVt2sQ/z/r1sGgRnHFG1eyJxeWX23M7/3w44ADo2TMz16krNGoEp55a/dfN6JrFqjoBmBCx7+awz2OBsZm0oa6xebMtX5rWpUtVYdIk+NOfKPj8QGAwHYcNgHo9YNOm8tfmzXD22XDJJbDHHrHP17w5jBxpL6DLUPh+FXF/bUHl2q1b7DItWyYuBKtXQ6tW0Y+JQK9eMG9e7O9HE4LAvu++i/29OXPsPVwIgmvefDPsvz/stde26+ymSjQRAOjTx96//BIOOST+OQJ7+/evuj2xOPdcu+cddsjcNZyq4YvX1zJ+8xurwB5+GI4/Pg0n/OgjuOEGW2C8e3eWHTOKRu8obUbfbR1600DXrvD55/HLJCIErVolJwTxztW7N7z+euzj8YTgq69if2/2bBOszp2jHz/iiNjfTReBEHzxReVCMGuWvWdSCMDWZHdqLvWybYCTON99B5Mnmzs/fLi9li5N4USlpTBhAhx9NBx0ECxcCI88AvPnU9CiFx07SlparAFdukBhIWzcGLtMfr6Ffzp1il0mGSFYsya2RwAmBIWFsGJF9ONLllgkbMcdt93fvbvZammtisyZY95AOp9fsuy4I7RrZ0JQGbNmWdkOHTJvl1NzcSGoRYwZY++ffw5/+5tFc/bcEx54wOr2SlmyxILU3brBr38NM2faiRYvhgsvhIYNWbYsfmWcCkGrOp5o5efbdeOFvBIVAtX4OQIwIYDY4aElS0zA6kX8h3TrZimSaAJSWmrhmMiwUHUjYjYEYZ94zJpl3kA2hcvJPi4EtQRVeO45OPBAS7hdc40lOw84AC67DPbbD959F374IUIUtmyBV16BoUOtOfvnP1stOHYsFBTYiZo23Vq8oAA6dkyv7YEQBOGWaCxZEjtRHJCoEPz8sz2DeELQq5e9x0oYx7InCDdFSxgvXmxeT7aFACw8NHdu/AbCpk0mhJkOCzk1HxeCWsKcOfD119v2KOjRwzrnPP+8tbYPP9wq8UaNoFOHUvbttIwR20/hht8uoGTufLjpJosvvfUWnHCCxT7CUCUjHkGXLvb+/fexy+Tnx4/pQ+JCsHp1eflYtG9vQpGqEERLGAct8H79Krcx0/TpY57LokWxy8ybByUlLgSOJ4trDc89Z2GT3/1u2/0icNJJFu7/4H1l2dTFLJv4Jcu+Xs8y3Yk5TfozjmMYPvZaBu6XF/caq1ZZKzHdHkHHjhZiieURbN5sApSIEBQX26tx49jlAiGI5xGImGMUTQiKi+HHH5P3CObMsTzHnnvGvm51EZ4w3m236GVmz7Z3FwLHhaAWUFZmrf6hQ6P0C1+5Ej77jBaffcZxr75qtVHLlnDZ2fCH/fmyeCf69IFvl+QxcL/411m2zN7T7RHUr29iEEsICgrsHhMRAoC1a61FH4vAa4jnEYAJwXPPmScUHiMPchnRhKBZM2jbNrYQ7L57fJGqLvbYw0Tpiy8qNh4CZs2y+/G+/Y4LQS1g6lSrpO+5B/P1X3sNPvsMpk8vj1GIwN57w2OPwSmnwHbbAdC9yA7H6/seUFBg7+n2CMAq1VihoUS6jkJ5xb5mTXwhSMQjAMsTrFtXMRwWq+toQLdusYXgoIPiX7O6aNzYRClewnjWLMtnRCbEndzDhaAW8NxzVq8f1/oj2PsYy4Z26WKds//wBxg4EPbZJ+oIo2bNrHvgt99Wfp1MeQRglerHH0c/VlnFG9Cypb1XlicIjlcmBEHPoblzkxeCYARxwOrV5knUhERxQJ8+8Mkn0Y+VlZlInHVWtZrk1FC8LVDD2bTJOvgM33cZTYcdac31RYustnr5Zbj6ahs1FGuYKdZZKFGPoF69+K3tVOnSxc4frRdLfr5dtzIBCvcI4pFIshhi9xxasiS+PYFHUFZWvi/os1/ThGDJEvN6Ilm0CIqKPD/gGC4EaaSsLMH+/Enw1lsWEz/1gwss8Dt1Kuy8c1Ln6NEjcY+gffs0T18RomtX66GyfHnFY/n5pm8RnZgqkIwQNGoETZrEL9emjd1v5FiCJUtsgFWDBtG/1727CfRPP5XvCxKvNU0IoKL3AtU3otipHbgQpJErr7R/rFijTlPhuTvzaccKjtj3Zxso0K5d0ufo0cPi8yUl8ctlYgxBQNCFNFrCOJGuo5C4EASDyRIZJBWt51BlYxqi9RyaM8fm0smEN5Uq4T2HIpk1y4Qu8Iqc3MaFIE1s2ABPPGGtr6B1WFXW//1R/jujPSd2/oT6kyeUB8mTpHt3E4EgGRyLTIwhCAgq1mgJ40SFINEcQbwJ5yLp3ds8gvAwT6pCUJO8ATBRb906esJ41iwTgcq8MCc3cCFIE2PHWswV4NVXq3gyVfjzn3n12v+xicac+uwxW3sBpUKPHvZeWZ4gGx5BSYkJUGWJYrAWbLNmiXsEidC7N/zyS/mzKS215xDPnuBY8J0tW0xMapoQiJhXEOkRqJZPLeE44EKQNkaPtv7YBx9cRSFYu9ZG/d58M8+1v4oePZR9D6pas617d3uPlycoKrKkYqY8gmbNrHKO9AiCBHIiHgEkNro4GY8gCI0EeYIffjBxiicE221nYaDAI1iwwAbF1YQRxZH06WNearjH88MPNuGeC4ET4EKQBhYtshzuOefAiBE2TfGCBSmcaNYs6wb63/+y/JZHeXdFL045peozgXbubIOL4nkEQdfRTHkEYJVrpEeQ6BiCgFatTCvjsXp14h5BMAo4yBMk2pU1fCxBrDUIagJ9+ljYMvxv74liJxIXgjTw1FPW3fCMM2xqaIBx45I4gaoNBBs82LqjfPABL7a8gLIySctqRfXrW2gmnkeQyTEEAekSgnSGhrbf3uyqqhA0bBh7KodsEiSMw/MEs2aVz1DqOOBCUGVKS00Ihg611nTnzja+K2EhKCoyBbngAhsPMKBBNfsAACAASURBVGsW7L8/Y8bYQOHdd0+PnZV1Ic3kqOKALl2sog3vVZWfb5VSrIVcIqlslbItW+yRJhoagm17DgVCEOQ0YtGtm5UNBmb16hW7u2k26dXLGinheYJZsyyMGWfoiZNjuBBUkSlTrDV9zjnl+4YPt9kfKuulw3vv2ejgMWNseugJE6BdO5YtsxkkTjwxfXZWNqisukJDRUXbhnaCPvuJ9l6pzCNIdFRxOL16WShvyxazp23bynPz3bpZXmD58prZYyigaVPYZZeKQuBhISccF4IqMnq0VRy/+U35vhEj7H38+BhfmjcPjj3WFn7fsMGWHbvxxq2TvkyaZMWOOSZ9dvboYYupBD2bIikosEo2bGmCtBNtOupEu44GVCYEic4zFE7v3lapBwO2E+nBFCTgp02zgWU1VQhg255Da9bYM3chcMJxIagCK1daZX/aadu2aHfbzZKQFXoPLV8O559v/5kffWSrgy1YYAsJhDFxorXM0znYJ6i4ok2WBpkdQxAQbYGaVIRgwwZrvUcj0eklwgmfc+j77xMTgsDm116z95ouBIsXWyPAp552ouFCUAXGjLEKKTwsFDB8OHzwgYkFxcVw660WmH36abjkEvvPvOaaCnMWl5RYuGno0PQuHxiMJYiVJ8jkGIKASCEoKbGJ2pIVAojtFaQSGtp9d3PG5s5N3CMIyrzxhr3XdCEA60bqPYacaLgQpIgqPPkkDBgAe+1V8fiIEZZI/O9/gd//Hm67zdYJ/vpruP/+KAsLGNOnWwx96ND02ht4BLHyBAUFmfcI2rUz3QtCQ8uWWbI9kYo3oDIhSMUjaNLEpm/64ANbajIRe5o0sUXiV6+255aM8FQ3gUh98YUJQYcONg7CcQJcCFJk1iz7xzr77OjH+/e3CuXVBwvMC7jpJnjppUonjHvrLevzf8QR6bW3bVsb1BXNI9i82fIHmfYIRMp7DkH5ezo9glRyBGDhoY8+ss+JClNgd00cSBZOly7WTTYQAvcGnEhcCFJk9Gib4fLkk6MfF4Hhh65hyuy2/Dz4KLj55oTOO3GiLUSf4rRCMRGJ3XNo+XLzcDLtEYBVSoFHkOwYAth2lbJoBAKR7PPr3bt85thEhSDwsmpyWAjKp5qYNg3mz3chcCriQpACxcW2WMyIEXFCEJs2MeKTq9hEY9467bmE5nZesQJmzEh/WCgg1liC6ug6GhA+qCwQgsr67IdT2cRzq1dDixbmVSVDeGI+WY+gpgsBmBDMnGli50LgROJCkALjx1uLNFqSeCvXXcf+3zxFuxabeHVq24TOO2WKvWdSCL77ruI02cF4h+rwCLp2tYXhi4tNCDp0MM8qURJJFqcSrw96DjVrlnh+IRhJvPfeyV+vugkXKxcCJxJfqjIGZWWW243Wgl6xwlqxhx0W48tvvAH330/eJZdwfHEjnn/eKr7KFjWfONFi+ZmqWLp3t2ToihWW6AyoTo8gaP0XFJgQJJMohsRyBKkIwS672Mjgrl0T7611yim2VlCS6wRlhaDnUMuWyYXinNzAhSAGhYVWMQ8cGP0f/eSTYyz6vWyZLQTbrx/8/e+MeB8efxzeeceEJRZlZTaQbMiQzC0mHj4ddbgQFBRYL5hketqkSngX0iVLYN99k/t+sPJYPCFI5T4aNrTKMplKsmHD5O3PFoHH069ferslO3UDF4IYFBba+1VXwciRCX6ptBROPdWa/y+8AI0bc9hh1mNj3Lj4QjBrll0zU2Eh2HY66v32K9+/bJl5A9VRQQQewXffWdI4lWk04o0uXrMmuZxDOK+9VncXamnWzHJaEWMXHQfIcI5ARIaKyAIRWSQi10U53kVE3hORWSLyhYikcVKFqhEIQVIrQ/7jH9YZ/Z//3BpAbtjQZpN47bXYo2HBvA+Ao45Kzd5ECFq7keGu6hhDENCpkwnOp5/agLJUwhTxhCBVjwBMDFNYCbTW8MorcOGF2bbCqYlkTAhEJA94BDga2BM4WUT2jCh2I/CSqvYHTgL+mSl7kiVpIdi0yYTg6KNtNtEwzjjDRhj/5S+xvz5xoi1FkMmBPk2b2pq6kV1IA4+gOmjY0BLEH3xg2+kUAtXUk8WOk8tk0iMYBCxS1W9VdTPwAjAsoowC24c+twB+yKA9SbFypb0nLAQvvGBZ2CuuqHBoyBCbj+iOO6wLXyRr11oLOZNhoYDILqRlZdUzz1A4XbrYDBuQfLIYYgtBUZF5GS4EjpMcmRSCjsDSsO2C0L5wbgVOE5ECYAJwSQbtSYrAI0ioUlGFBx6wzugxgrAPPmgJ2jPOsBRCOO+8Y+mF6hCCyEFlK1dayKq6PALYtvJPJZ4fa5WyVKaXcBwn++MITgaeUtVOwDHAMyJSwSYRGSUiM0RkRmFQQ2eYwkKrUBJabOSjjyzb+8c/xsy4tmplcxN99ZXNOB3OxIk2CCo8gZspevSwid6CfEV1jiEICCr/9u2tB1CyxFqcJpUJ5xzHyawQLAPC153qFNoXzrnASwCq+inQGKgw+kpVH1PVAao6oF01ZfMKC5MICz3wgNX0p50Wt9iQITb/3L332hrHYM7ExIk2t1ACg4+rTPfuFg4Kn/gNsuMRpNqfvVUrWL++fEqIAPcIHCc1MikEnwG7iEh3EWmIJYNfjyjzPXA4gIjsgQlB9TT5KyFhIViyxPqGnn9+Qqu63H23VcZnnQU//2weQkFB9YSFoOJ01NnwCNIhBFAxPOQegeOkRsaEQFVLgIuBScDXWO+geSJyu4gcFyp2JXC+iMwBngfOUo2cACE7JCwEjzxi4aCLLkrovM2a2WSk+flw9dXl3UaHDEnZ1KSInI562TKblyd8gFmmCUJDqSSKIfbo4lRnHnWcXCejwQhVnYAlgcP33Rz2+SvggEzakCqFhTB4cCWFNmywYcPDhyeV9TzwQBuodvfd1hLv1SvxxdurSseOlvcI9wh22in5Sdqqws472+M6IMW/fGVC4KEhx0kOH1kchbIy601TqUfw7LMWn7j00qSvcfvttlb9vHnpXaS+MvLyrCUe7hFUZ34ALIIWvlxlssQSgjVrbJxCJtdddpy6SLZ7DdVI1q61RGRcIVC1PqF7751S07ZxY/i//7NKuDqFALYdS1Cdo4rTRTyPoHVrn0vHcZLFPYIoJDSq+O23LdP79NMp1zx7712erK1OevSwdQ/APIJMTmuRCeJ5BB4WcpzkcY8gCgkJwQMPWIa1upvzaaB7d2s9FxRYz6Xa6hFE9hpKdQpqx8l1XAiiUKkQLFwIb75pgwKSWVWlhhB0If3wQ3uv7hxBVWnc2HIB0UJD7hE4TvK4EEShUiF46CHrevP731ebTekk6EIaCEFt8whEos835BPOOU5quBBEIRCCttFWmNy40fICI0faHAm1kNruEUB0IfDQkOOkhgtBFFautIFfUZeWfOUVm9/g/POr3a500aqVzW00d65t1wUh2LLF8h0eGnKc5HEhiELcUcWjR9uIqF/9qlptSjeBV9CmTeVrKddEIoUgSBy7R+A4yeNCEIWYQrB4Mbz/PpxzTq3vrB7kCWpbfiAgUgh8egnHSR0XgijEFIL//MdWlj/zzGq3Kd0EHkFtDAtBRSEIPntoyHGSx4UgClGFoLQUnnrKpgmtrbVnGHXBI1i3zqYDAfcIHKcquBBEoBpDCCZNsmG4556bFbvSTW33CFq2tL/VunW27RPOOU7quBBEUFRk69BXEILRo23nscdmxa50s9tu9t6zZ3btSJXI0cW+FoHjpI4LQQRRB5MVFsLrr8Ppp9uQ1jpA9+4239DIkdm2JDUi5xsKPIKWLbNjj+PUZnzSuQiiCsGzz1pH9XPOyYpNmWKffbJtQepECsGaNbD99tWz3Kfj1DXcI4igghCo2qrz++5rK8g4NYJoHoGHhRwnNSoVAhH5jYjkjGBUEILp0231mDrmDdR2XAgcJ30kUsGfCCwUkb+LyO6ZNijbVJhnaPRoW/LqpJOyZpNTkWihIe8x5DipUakQqOppQH9gMfCUiHwqIqNEpHnGrcsChYU2s3SzZtiaxM8/D7/7nQWgnRrDdttZPsA9AsepOgmFfFR1PTAWeAHYCRgOfC4il2TQtqwQjCEQwSaY+/lnDwvVQCKnonaPwHFSJ5EcwXEiMg54H2gADFLVo4G+wJWZNa/62WbR+ieftI72Bx2UVZuc6LRsaQKg6h6B41SFRDrbnQDcp6pTw3eq6kYRqRvDbMPYOqq4sBCmToXbb6/1E8zVVVq1sgFlGzZY714XAsdJjURCQ7cC04MNEWkiIt0AVPWdjFiVRbYKwfTQLR98cFbtcWIThIZ8wjnHqRqJCMHLQFnYdmloX51kqxBMmwZ5ebV71FUdJxACn3DOcapGIkJQX1U3Bxuhz3VjnoUIiottrqGtQtC7t3VPcWok7hE4TnpIRAgKReS4YENEhgErM2dS9tg6mKxNmYWGBg3KrkFOXIIcwapVtu0egeOkRiLJ4t8Dz4nIw4AAS4EzMmpVltgqBCXLrYbZd9/sGuTEpVUrWybi++9t24XAcVKjUiFQ1cXAfiLSLLRdlHGrssRWIVj+hX1wIajRBKGgxYu33XYcJzkSmqtRRH4N9AIaS6grparenkG7ssJWIcj/zIYW77FHdg1y4hJU/N9+Cw0aeDrHcVIlkQFlj2LzDV2ChYZ+B3TNsF1ZYasQzHsfBg60XkNOjSVYe2DxYgsL+XAPx0mNRJLF+6vqGcAaVb0NGAzsmsjJRWSoiCwQkUUicl2U4/eJyOzQ6xsRWZuc+emlsBDy8pQWcz/2sFAtIPAI8vM9LOQ4VSGR0FBx6H2jiHQAVmHzDcVFRPKAR4AjgQLgMxF5XVW/Csqo6uVh5S/BJrfLGoWF0LbFFuqt3uxCUAsIKv/Nmz1R7DhVIRGP4L8i0hK4G/gcyAfGJPC9QcAiVf02NPbgBWBYnPInA88ncN6MUVgI7Rqttw3vOlrjCfcC3CNwnNSJ6xGEFqR5R1XXAq+IyBtAY1Vdl8C5O2JdTQMKgKjNbBHpCnQH3o1xfBQwCqBLly4JXDo1Vq6EdqU/QadO0KFDxq7jpIfmzaFePSgrc4/AcapCXI9AVcuw8E6wvSlBEUiWk4Cxqloaw47HVHWAqg5ot81iwumlsBDabfjOw0K1hHr1yhPGLgSOkzqJhIbeEZETRJLuk7EM6By23Sm0LxonkeWwEEDhijLabch3IahFBCEhDw05TuokIgQXYJPMbRKR9SLys4isT+B7nwG7iEh3EWmIVfavRxYKLX/ZCvg0CbvTzpYtsGZtPdpR6EJQiwgEwD0Cx0mdREYWp7QkpaqWiMjFwCQgDxitqvNE5HZghqoGonAS8IKqairXSRfBfDXtZKXPOFqLcCFwnKpTqRCIyK+i7Y9cqCZGmQnAhIh9N0ds31rZeaqDrYPJOjfxIaq1iCBH4KEhx0mdRMYRXB32uTHWLXQmcFhGLMoShT+VAfVot1f7bJviJIF7BI5TdRIJDf0mfFtEOgP3Z8yiLFH45Y9AB9oN6p5tU5wk8GSx41SdRJLFkRQAdW42tsLPbchDu4P3zLIlTjK0abPtu+M4yZNIjuAhIEjk1gP6YSOM6xSF8y1b3Hq/hKZRcmoIp59uItC2bbYtcZzaSyI5ghlhn0uA51X14wzZkzHy8+GDD+DMM6MfL1yykdb111G/UYtqtcupGu3bwznnZNsKx6ndJCIEY4HiYNSviOSJSFNV3ZhZ09LLiy/CddfBoYdChVkqiospXCm0a7UpK7Y5juNkk4RGFgNNwrabAG9nxpzMMXy4vY8fH+XgrFkUahva7ZhKysRxHKd2k0jN1zh8ecrQ56aZMykz7Lor9OoFr74a5eC0aRTSjnbdfPyA4zi5RyJCsEFE9g42RGQf4JfMmZQ5RoyADz8sHzy2lWnTWFlvBxtM5jiOk2MkIgSXAS+LyIci8hHwInBxZs3KDMOH25TFr0fMeFT2v+ms0tZkcGJTx3GcGksiA8o+C00Mt1to1wJV3ZJZszJDv37QrRuMGwfnnhvaWVjImvy1lJLnQuA4Tk6SyOL1FwHbqepcVZ0LNBORCzNvWvoRsfDQlCmwPpg/de5cCjEFcCFwHCcXSSQ0dH5ohTIAVHUNcH7mTMosw4fbGrcTgqnwVq50IXAcJ6dJRAjywhelCS1K3zBzJmWWwYNhxx0tPATAqlUuBI7j5DSJCMFE4EUROVxEDsdWEnsrs2Zljrw8GDbMPILiYlwIHMfJeRIRgmuxReV/H3p9ybYDzGodI0ZAURG8/TYmBA06Aj5fjeM4uUmlQhBawH4akI+tRXAY8HVmzcoshx4KLVqEBpetXk1ho040bw6NGmXbMsdxnOonZvdREdkVODn0WomNH0BVD60e0zJHw4Zw7LE2nqBk3zUU1t+Jdr6wieM4OUo8j2A+1vo/VlUPVNWHgNLqMSvzjBhh6xR/mN+Zwno7eH7AcZycJZ4QjACWA++JyOOhRLHEKV+rGDIEGjeGV3/Yl8KyNi4EjuPkLDGFQFXHq+pJwO7Ae9hUEzuIyL9E5KjqMjBTbLcdDB0K49YfzorNLV0IHMfJWRJJFm9Q1TGhtYs7AbOwnkS1nuHDylhW1oHlG10IHMfJXZKagF9V16jqY6p6eKYMqk6O/dV66mPTJrkQOI6Tq+T0Siyty1ZyKO8BLgSO4+QuOS0ErFrFcGyuCRcCx3FylZwXgtN4lhvOXMbBB2fbGMdxnOyQyOL1dZfVq2lOEXf+aSP4KpWO4+QoOe8RANCmTXbtcBzHySIuBCI28ZDjOE6O4kLQqpXNTe04jpOjZFQIRGSoiCwQkUUicl2MMiNF5CsRmSciYzJpTwVWrfKwkOM4OU/GksWhlcweAY4ECoDPROR1Vf0qrMwuwPXAAaq6RkR2yJQ9UVm92oXAcZycJ5MewSBgkap+q6qbgReAYRFlzgceCa2DjKquyKA9FXGPwHEcJ6NC0BFYGrZdENoXzq7AriLysYj8T0SGRjuRiIwSkRkiMqOwsDB9FroQOI7jZD1ZXB/YBTgEWwDncRFpGVkoNL/RAFUd0C6dQ4BXrYLWviKN4zi5TSaFYBnQOWy7U2hfOAXA66q6RVW/A77BhCHzbN5sCxe7R+A4To6TSSH4DNhFRLqLSEPgJOD1iDLjMW8AEWmLhYq+zaBN5axebe8uBI7j5DgZEwJVLQEuBiZhi92/pKrzROR2ETkuVGwSsEpEvsIWv7laVVdlyqZt8FHFjuM4QIbnGlLVCcCEiH03h31W4IrQq3pxIXAcxwGynyzOHoEQeLLYcZwcx4XAPQLHcXIcFwIXAsdxcpzcFYLVq6FRI2jaNNuWOI7jZJXcFYJgVLFIti1xHMfJKrktBJ4odhzHyXEh8PyA4ziOC4HjOE6uk7tC4GsROI7jALkqBKruETiO44TITSEoKoItW1wIHMdxyFUh8OklHMdxtpLbQuAegeM4To4Kga9F4DiOs5XcFAL3CBzHcbbiQuA4jpPj5LYQtGqVXTscx3FqALkrBNtvDw0aZNsSx3GcrJObQuCjih3HcbaSm0Lgo4odx3G24kLgOI6T47gQOI7j5Di5KwQ+vYTjOA6Qi0JQWgpr17pH4DiOEyL3hGDNGnt3IXAcxwFyUQh8VLHjOM42uBA4juPkOLkrBJ4sdhzHAXJRCHwKasdxnG2on20Dqh0PDTlOTrJlyxYKCgooLi7OtikZpXHjxnTq1IkGScylllEhEJGhwANAHvCEqt4Vcfws4G5gWWjXw6r6RCZtYtUqqF/fJp1zHCdnKCgooHnz5nTr1g0RybY5GUFVWbVqFQUFBXTv3j3h72UsNCQiecAjwNHAnsDJIrJnlKIvqmq/0CuzIgDlg8nq6A/BcZzoFBcX06ZNmzorAgAiQps2bZL2ejKZIxgELFLVb1V1M/ACMCyD10sMn17CcXKWuiwCAancYyaFoCOwNGy7ILQvkhNE5AsRGSsinaOdSERGicgMEZlRWFhYNat8egnHcZxtyHavof8C3VS1DzAFeDpaIVV9TFUHqOqAdu3aVe2KvhaB4zhZYO3atfzzn/9M+nvHHHMMa9euzYBF5WRSCJYB4S38TpQnhQFQ1VWquim0+QSwTwbtMTw05DhOFoglBCUlJXG/N2HCBFq2bJkps4DM9hr6DNhFRLpjAnAScEp4ARHZSVWXhzaPA77OoD2GC4HjOJddBrNnp/ec/frB/ffHPHzdddexePFi+vXrR4MGDWjcuDGtWrVi/vz5fPPNNxx//PEsXbqU4uJiLr30UkaNGgVAt27dmDFjBkVFRRx99NEceOCBfPLJJ3Ts2JHXXnuNJk2aVNn0jHkEqloCXAxMwir4l1R1nojcLiLHhYr9UUTmicgc4I/AWZmyB4BffrGXC4HjONXMXXfdxc4778zs2bO5++67+fzzz3nggQf45ptvABg9ejQzZ85kxowZPPjgg6wKxjyFsXDhQi666CLmzZtHy5YteeWVV9JiW0bHEajqBGBCxL6bwz5fD1yfSRu2waeXcBwH4rbcq4tBgwZt09f/wQcfZNy4cQAsXbqUhQsX0iai0dq9e3f69esHwD777EN+fn5abMmtkcU+vYTjODWE7bbbbuvn999/n7fffptPP/2Upk2bcsghh0QdC9CoUaOtn/Py8vjll1/SYku2ew1VLz69hOM4WaJ58+b8/PPPUY+tW7eOVq1a0bRpU+bPn8///ve/arUttzwCFwLHcbJEmzZtOOCAA+jduzdNmjRhxx133Hps6NChPProo+yxxx7stttu7LffftVqmwuB4zhONTFmzJio+xs1asRbb70V9ViQB2jbti1z587duv+qq65Km125GRryZLHjOM5WcksIVq+Gpk2hceNsW+I4jlNjyC0h8MFkjuM4FXAhcBzHyXFcCBzHcXIcFwLHcZwcJ7eEYPVq7zHkOE5WSHUaaoD777+fjRs3ptmicnJHCFR9LQLHcbJGTRaC3BlQtm4dlJa6EDiOk41ZqLeZhvrII49khx124KWXXmLTpk0MHz6c2267jQ0bNjBy5EgKCgooLS3lpptu4qeffuKHH37g0EMPpW3btrz33nvpNZxcEgIfVew4Tha56667mDt3LrNnz2by5MmMHTuW6dOno6ocd9xxTJ06lcLCQjp06MCbb74J2BxELVq04N577+W9996jbdu2GbHNhcBxnJwj27NQT548mcmTJ9O/f38AioqKWLhwIQcddBBXXnkl1157LcceeywHHXRQtdiTO0LgU1A7jlNDUFWuv/56LrjgggrHPv/8cyZMmMCNN97I4Ycfzs033xzlDOkld5LFPs+Q4zhZJHwa6iFDhjB69GiKiooAWLZsGStWrOCHH36gadOmnHbaaVx99dV8/vnnFb6bCXLHI/DQkOM4WSR8Guqjjz6aU045hcGDBwPQrFkznn32WRYtWsTVV19NvXr1aNCgAf/6178AGDVqFEOHDqVDhw4ZSRaLqqb9pJlkwIABOmPGjOS/+Npr8NRTMHYs5OWl3S7HcWo2X3/9NXvssUe2zagWot2riMxU1QHRyueORzBsmL0cx3GcbcidHIHjOI4TFRcCx3FyhtoWCk+FVO7RhcBxnJygcePGrFq1qk6LgaqyatUqGie5+Fbu5Agcx8lpOnXqREFBAYWFhdk2JaM0btyYTp06JfUdFwLHcXKCBg0a0L1792ybUSPx0JDjOE6O40LgOI6T47gQOI7j5Di1bmSxiBQCS1L8eltgZRrNqS3k6n1D7t6733dukch9d1XVdtEO1DohqAoiMiPWEOu6TK7eN+Tuvft95xZVvW8PDTmO4+Q4LgSO4zg5Tq4JwWPZNiBL5Op9Q+7eu993blGl+86pHIHjOI5TkVzzCBzHcZwIXAgcx3FynJwRAhEZKiILRGSRiFyXbXsyhYiMFpEVIjI3bF9rEZkiIgtD762yaWMmEJHOIvKeiHwlIvNE5NLQ/jp97yLSWESmi8ic0H3fFtrfXUSmhX7vL4pIw2zbmglEJE9EZonIG6HtOn/fIpIvIl+KyGwRmRHaV6XfeU4IgYjkAY8ARwN7AieLyJ7ZtSpjPAUMjdh3HfCOqu4CvBParmuUAFeq6p7AfsBFob9xXb/3TcBhqtoX6AcMFZH9gL8B96lqT2ANcG4WbcwklwJfh23nyn0fqqr9wsYOVOl3nhNCAAwCFqnqt6q6GXgBqJPrVqrqVGB1xO5hwNOhz08Dx1erUdWAqi5X1c9Dn3/GKoeO1PF7V6MotNkg9FLgMGBsaH+du28AEekE/Bp4IrQt5MB9x6BKv/NcEYKOwNKw7YLQvlxhR1VdHvr8I7BjNo3JNCLSDegPTCMH7j0UHpkNrACmAIuBtapaEipSV3/v9wPXAGWh7Tbkxn0rMFlEZorIqNC+Kv3OfT2CHENVVUTqbJ9hEWkGvAJcpqrrrZFo1NV7V9VSoJ+ItATGAbtn2aSMIyLHAitUdaaIHJJte6qZA1V1mYjsAEwRkfnhB1P5neeKR7AM6By23Sm0L1f4SUR2Agi9r8iyPRlBRBpgIvCcqr4a2p0T9w6gqmuB94DBQEsRCRp6dfH3fgBwnIjkY6Hew4AHqPv3jaouC72vwIR/EFX8neeKEHwG7BLqUdAQOAl4Pcs2VSevA2eGPp8JvJZFWzJCKD78JPC1qt4bdqhO37uItAt5AohIE+BILD/yHvDbULE6d9+qer2qdlLVbtj/87uqeip1/L5FZDsRaR58Bo4C5lLF33nOjCwWkWOwmGIeMFpV78yySRlBRJ4HDsGmpf0JuAUYD7wEdMGm8B6pqpEJ5VqNiBwIfAh8SXnM+AYsT1Bn711E+mDJwTysYfeSqt4uIj2wlnJrYBZwmqpuyp6lmSMUGrpKVY+t6/cdur9xoc36wBhVvVNE2lCF33nOCIHjOI4TnVwJDTmO4zgxcCFwHMfJcVwIHMdxchwXAsdxnBzHhcBxHCfHcSFwT2YgDwAAAZVJREFUnAhEpDQ0s2PwSttEdSLSLXxmWMepCfgUE45TkV9UtV+2jXCc6sI9AsdJkNA88H8PzQU/XUR6hvZ3E5F3ReQLEXlHRLqE9u8oIuNCawXMEZH9Q6fKE5HHQ+sHTA6NCHacrOFC4DgVaRIRGjox7Ng6Vd0LeBgbqQ7wEPC0qvYBngMeDO1/EPggtFbA3sC80P5dgEdUtRewFjghw/fjOHHxkcWOE4GIFKlqsyj787FFYL4NTXD3o6q2EZGVwE6quiW0f7mqthWRQqBT+BQHoSmyp4QWEEFErgUaqOodmb8zx4mOewSOkxwa43MyhM99U4rn6pws40LgOMlxYtj7p6HPn2AzYAKcik1+B7Zk4B9g6+IxLarLSMdJBm+JOE5FmoRW/AqYqKpBF9JWIvIF1qo/ObTvEuA/InI1UAicHdp/KfCYiJyLtfz/ACzHcWoYniNwnAQJ5QgGqOrKbNviOOnEQ0OO4zg5jnsEjuM4OY57BI7jODmOC4HjOE6O40LgOI6T47gQOI7j5DguBI7jODnO/wNLZPqLNGyt3wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the accuracy and loss curves\n",
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(0,50)\n",
    "\n",
    "plt.plot(epochs, acc, 'r')\n",
    "plt.plot(epochs, val_acc, 'b')\n",
    "plt.title('Training and Testing accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train', 'test'], loc='lower right')\n",
    "plt.figure()\n",
    "#------------------------------------------------\n",
    "# Plot training and validation loss per epoch\n",
    "#------------------------------------------------\n",
    "plt.plot(epochs, loss, 'r')\n",
    "plt.plot(epochs, val_loss, 'b')\n",
    "plt.legend(['train', 'test'], loc='upper right')\n",
    "plt.title('Training and Testing Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.figure()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cGdce6Pcyn0s"
   },
   "source": [
    "## Validation of the trained model\n",
    "Show the errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 402
    },
    "id": "39ih5u62yn0s",
    "outputId": "7eae79ac-a09b-48d4-bfec-57a50a1f6aea",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 604 images belonging to 10 classes.\n",
      "<class 'dict'>\n",
      "{0: 'airplanes', 1: 'bonsai', 2: 'butterfly', 3: 'car_side', 4: 'faces', 5: 'ketch', 6: 'leopards', 7: 'motorbikes', 8: 'scorpion', 9: 'watch'}\n",
      "19/19 [==============================] - 10s 529ms/step\n",
      "No of errors = 55/604\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a generator for prediction\n",
    "validation_generator = validation_datagen.flow_from_directory(\n",
    "        validation_dir,\n",
    "        target_size=(image_size, image_size),\n",
    "        batch_size=val_batchsize,\n",
    "        class_mode='categorical',\n",
    "        shuffle=False)\n",
    "\n",
    "# Get the filenames from the generator\n",
    "fnames = validation_generator.filenames\n",
    "\n",
    "# Get the ground truth from generator\n",
    "# ground_truth = validation_generator.classes\n",
    "\n",
    "ground_truth = validation_generator.classes\n",
    "label2index = validation_generator.class_indices\n",
    "idx2label = dict((v,k) for k,v in label2index.items())\n",
    "# # Get the label to class mapping from the generator\n",
    "# print(\"idx2label\")\n",
    "print(type(idx2label))\n",
    "print(idx2label)\n",
    "# Get the predictions from the model using the generator\n",
    "predictions = cnn.predict(validation_generator,verbose=1)\n",
    "predicted_classes = np.argmax(predictions,axis=1)\n",
    "\n",
    "errors = np.where(predicted_classes != ground_truth)[0]\n",
    "print(\"No of errors = {}/{}\".format(len(errors),validation_generator.samples))\n",
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "def PlotConfusionMatrix(ground_truth, predicted_classes):\n",
    "    np.set_printoptions(precision=2)\n",
    "    classes = ['airplanes','bonsai','butterfly','car_side','faces','ketch','leopards','motorbikes','scorpion','watch']\n",
    "    title='Confusion matrix'\n",
    "    cmap=plt.cm.Blues\n",
    "    cm = confusion_matrix(ground_truth,predicted_classes)\n",
    "    fig, ax = plt.subplots()\n",
    "    im = ax.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    ax.figure.colorbar(im, ax=ax)\n",
    "    ax.set(xticks=np.arange(cm.shape[1]),\n",
    "           yticks=np.arange(cm.shape[0]),\n",
    "           xticklabels=classes, yticklabels=classes,\n",
    "           title=title,\n",
    "           ylabel='True label',\n",
    "           xlabel='Predicted label')\n",
    "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\",rotation_mode=\"anchor\")\n",
    "    fmt ='d' \n",
    "    thresh = cm.max() / 2.\n",
    "    for i in range(cm.shape[0]):\n",
    "        for j in range(cm.shape[1]):\n",
    "            ax.text(j, i, format(cm[i, j],fmt),\n",
    "                    ha=\"center\", va=\"center\",\n",
    "                    color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "    fig.tight_layout()    \n",
    "    plt.show()\n",
    "PlotConfusionMatrix(ground_truth, predicted_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "u9PqvEqUyn0t",
    "outputId": "d74c7d2e-2289-46a4-9725-a3b0d7fd05ae"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training and validation done.\n"
     ]
    }
   ],
   "source": [
    "print(\"Training and validation done.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "lVdsTZFwyn0u"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "keras-finetune-vgg16.ipynb",
   "provenance": []
  },
  "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}