solution_project1_German Traffic Sign.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "O9jAqShELHZI"
   },
   "source": [
    "![alt text](https://drive.google.com/uc?export=view&id=1UXScsVx_Wni_JuDdB8LeTnM6jsPfIwkW)\n",
    "\n",
    "Proprietary content. © Great Learning. All Rights Reserved. Unauthorized use or distribution prohibited."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SRkFT8HtCZHS"
   },
   "source": [
    "# German Traffic Sign Recognition\n",
    "Multi-class, single-image classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SJWdbMALTcGJ"
   },
   "source": [
    "### Dataset\n",
    "The German Traffic Sign Benchmark is a multi-class, single-image classification challenge held at the International Joint Conference on Neural Networks (IJCNN) 2011. They cordially invite researchers from relevant fields to participate: The competition is designed to allow for participation without special domain knowledge. Their benchmark has the following properties:\n",
    "\n",
    "- Single-image, multi-class classification problem\n",
    "- More than 40 classes\n",
    "- More than 50,000 images in total\n",
    "- Large, lifelike database"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "hRhJulWyIGMN"
   },
   "source": [
    "#### Notes\n",
    "- For this project, we have reduced the number of images. There are around 16,500+ images in the dataset provided. You can check “label_details” folder for getting information about the classes.\n",
    "- If the model is taking too much time to get trained then you can reduce the number of classes. There are around 43 classes in the dataset, model should be trained on a minimum of 15 classes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "tf.__version__\n",
    "import numpy as np\n",
    "import cv2\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from glob import glob\n",
    "import os\n",
    "# Ignore the warnings\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_path = 'Train-20210114T110605Z-001/Train'\n",
    "number_of_classes = 15"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qC37j3FUDVYd"
   },
   "source": [
    "### Initialize ImageDataGenerator (7 Marks)\n",
    "- Rescale the images\n",
    "- Specify value for validation_split & get 75% data in training and 25% data in training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ImageDataGenerator(\n",
    "    featurewise_center=False, samplewise_center=False,\n",
    "    featurewise_std_normalization=False, samplewise_std_normalization=False,\n",
    "    zca_whitening=False, zca_epsilon=1e-06, rotation_range=0, width_shift_range=0.0,\n",
    "    height_shift_range=0.0, brightness_range=None, shear_range=0.0, zoom_range=0.0,\n",
    "    channel_shift_range=0.0, fill_mode='nearest', cval=0.0,\n",
    "    horizontal_flip=False, vertical_flip=False, rescale=None,\n",
    "    preprocessing_function=None, data_format=None,\n",
    "    validation_split=0.25, dtype=None\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "IMG_SIZE = 28"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# using rescale to image value range :[0,1]\n",
    "# Training :validation = 75: 25\n",
    "image_datagen = tf.keras.preprocessing.image.ImageDataGenerator(\n",
    " rescale=1.0/255.0,\n",
    " validation_split=0.25\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "58e4ODapEVdx"
   },
   "source": [
    "### Get training data from ImageDataGenerator (5 Marks)\n",
    "- Give directory path\n",
    "- Give target size\n",
    "- Give batch_size\n",
    "- Specify classes, if you wish to use less number of classes you need to give class names in a list (Atleast 15 classes should be there)\n",
    "- Specify class_mode\n",
    "- Specify color_mode\n",
    "- Specify subset\n",
    "\n",
    "You can get details here\n",
    "https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/image/ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GSfT2cB4COgC"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 16248 images belonging to 15 classes.\n"
     ]
    }
   ],
   "source": [
    "train_generator = image_datagen.flow_from_directory(\n",
    "        directory = dataset_path,\n",
    "        target_size=(IMG_SIZE, IMG_SIZE),\n",
    "        batch_size=32,\n",
    "        classes= ['0','1','2','3','4','5','6','7','8','9','11','12','13','14','15'],\n",
    "        class_mode = 'categorical',\n",
    "        color_mode=\"rgb\",\n",
    "        subset='training',\n",
    "        shuffle=True,\n",
    "        seed =1\n",
    "        )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "GzYNcoV9Gawj"
   },
   "source": [
    "### Get validation data from ImageDataGenerator (5 Marks)\n",
    "- Give directory path\n",
    "- Give target size\n",
    "- Give batch_size\n",
    "- Specify classes, if you wish to use less number of classes you need to give class names in a list (Atleast 15 classes should be there)\n",
    "- Specify class_mode\n",
    "- Specify color_mode\n",
    "- Specify subset\n",
    "\n",
    "You can get details here\n",
    "https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/image/ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-m9zfw0AHIma"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 5412 images belonging to 15 classes.\n"
     ]
    }
   ],
   "source": [
    "validation_generator = image_datagen.flow_from_directory(\n",
    "        directory = dataset_path,\n",
    "        target_size=(IMG_SIZE, IMG_SIZE),\n",
    "        batch_size=32,\n",
    "        classes= ['0','1','2','3','4','5','6','7','8','9','11','12','13','14','15'],\n",
    "        class_mode = 'categorical',\n",
    "        color_mode=\"rgb\",\n",
    "        subset='validation',\n",
    "        shuffle=True,\n",
    "        seed =1\n",
    "        )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize few images before giving to model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_batch, y_batch = next(train_generator)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAAZuElEQVR4nO2da3Ccd3XGn7O7WkmW5Yt8kWzHjp04IRcwDtGESygNtzTJl8C0dMiHks4wNTOFFjpMpwz9QIZPGabA8KGlYyBD2kIYOiElM6QQN4Wm0BLiJMaOL4kd2/Fd8k13abX7vqcftKEm6P/8hS67Gv7Pb0YjaR+d9/3vu/vsu9rznnPM3SGE+N2n0OwFCCEag8wuRCLI7EIkgswuRCLI7EIkQqmRO+tcvtrX9GwK6rHEgFlYi2UVCsb1PLJvplteobFZFnlNLeRctjLVS8Xw9nPMNdsSO27kQQFgNJ5v26xI9QLfdQS+72iWij0ZZ7R79pjzbbOVXeg7ieHBS9NuYE5mN7O7AHwFQBHA1939Qfb3a3o24fP/8NOg7vQAAFYKHwSv1WhsazGjeqXGH9zhybDWNvoKjR0YbqW6t49TvbN1C9WXd7YHtcmcLByAFyNmzflxHa/xF7IWhB9Tz6s0ttTaQfVWfljBXt8tYuZqja/NIsct94i1amNBycFf5DKy9gf+8u6gNuu38Tb1svv3AO4GcBOA+8zsptluTwixsMzlf/bbABxx96PuPgngOwDunZ9lCSHmm7mYfQOAk1f8fqp+269hZjvMbLeZ7R4auDCH3Qkh5sJczD7dPy2/8c+Eu+9091537122YvUcdieEmAtzMfspABuv+P0qAGfmthwhxEIxF7M/C+A6M9tiZmUAHwbw+PwsSwgx38w69ebuNTP7BIAfYSr19pC772cx1clJ9J08HtTHnL8x6FmzOaiVCzwXPeGR17XJY1T+4WOPBDUf4umtU71vp/pfve0PqW4Zz+OPTA4Htarz1FmxEsnpRtLJ1Ui6NMtJGinykFQqE3zfkzydWmppI7E8tebgj2mxyK3jWTi1NkX4zhfB71ehRA4cScvNKc/u7k8AeGIu2xBCNAZdLitEIsjsQiSCzC5EIsjsQiSCzC5EIsjsQiRCQ+vZiy0ldK0LXzK7PO+i8Z6Hc7rjNZ6LtkhNeC1S4rq2M1xPeRHLaexHNmynep5FylBrPJc9UQnf9xZSFgwAXuTHJavy41rgKWFkLGccic0L/BqBWo3n4ctt4fgsVk4daXCQZZH4SAkte1TyiCutGn6+sDp8ndmFSASZXYhEkNmFSASZXYhEkNmFSASZXYhEaGjqzd0xyUoLI+15vRbO1RQjJa5V0s0TAAotV1H91vf9RVCzQqTtcCTFNDIULlEFAHe+ASuEX7OrPEOEbHSAb7sUeYqQNtYAULCwbhl/vMsZL0OtRJ6+lZGhoHbsDK3Gxt5nfkH1ibP9VO+44c1Uf/edHwlqyzBCY2mFK4nTmV2IRJDZhUgEmV2IRJDZhUgEmV2IRJDZhUgEmV2IRGhont3cUSClhRkiuXLS3tczXg4ZqWiER0Y6F/JwrrtS5bGtZX6YC6RkEQCsLVKeOxm+717l+/7BD79J9YlIfJ7xawCWtofXfv7wYRrb3jJA9cmLPNdduRiOz6p8ci7I9FkAKETGKmfjR6l+pP1NQW37e66nsT5Brj9QiasQQmYXIhFkdiESQWYXIhFkdiESQWYXIhFkdiESoaF5dgA0fengrYGRkXx2gec9rcjzwR6pra7l4dfFUinSNpjEAkBeIGONAdTG+far4+Hj9pN//0cae+oAr9te2c7bZF86+yrV+yqjQa2FtAYHgPFoO+aIzvojxFpJR86DHsvDd66nek/vpvC+a7yOPy+1EDV8n+dkdjM7DmAYU+0Zau7eO5ftCSEWjvk4s7/b3S/Mw3aEEAuI/mcXIhHmanYH8KSZPWdmO6b7AzPbYWa7zWz38NClOe5OCDFb5vo2/nZ3P2NmawHsMrND7v70lX/g7jsB7ASAzVu3RTozCiEWijmd2d39TP17P4DHANw2H4sSQsw/sza7mXWYWedrPwO4E8CL87UwIcT8Mpe38d0AHqvnMksAvu3uP2QB1do4zp7fF9SXd95Id9jSEV5uIeOvWxnpOQ8ALZH+5xOT4Vz2cz97hsYe+PnXI/teRfXzx1+guk+Ee+Kz/gEAH4MNAOcj54Mi6VkPAOwKgsiYAEwW+R+0LltB9fXrbwpq/YefpbHVCu+PkLXxxww33kLl7rbwfXPjtqzlpH8BufZg1mZ396MAeCd8IcSiQak3IRJBZhciEWR2IRJBZhciEWR2IRKhsa2kUUIL1gT1sQpvDdzZujaoVVl7XQA//69HqX5y//NUL9bOBLVLR0/R2EKNl+6O+SGqlyJti52k1yLZLSCSOssjo7CXb9xC9SXdq8PbLrbR2ILxfXvkzo2TFFWVpFIBAMbLjldey0tY73jbh6jOLiVl6TMAaGFjsMkjrjO7EIkgswuRCDK7EIkgswuRCDK7EIkgswuRCDK7EInQ2FbSDhSq4dxnpbyEho8NnAtqv3h42q5Yv+LwizyHHysFzUlbY4/lwSNth73UTvWurddRvWVluNyyUOygsXnO156DX7+AyMhmFFnbYx6bRVpJRxPt/cfDoR4paW4NXx8AAEsmeqje3crvm+XkvkVqf70Qti3L0evMLkQiyOxCJILMLkQiyOxCJILMLkQiyOxCJILMLkQiNDTPPlGt4KWzrwT1bddvo/HHDoXb/3Zv5gNkX973A6pnkZxtaenSoLbp5rfSWLTyPHolmk6ODdIJ6znL5wIoFCKjiyN13bVInp3lswuR6w9y59suRp6+Z44dDGoGXiu/eStva95x9wep3lbm20eRtEW3yDho0v67QHL0OrMLkQgyuxCJILMLkQgyuxCJILMLkQgyuxCJILMLkQgNzbMX4egshvO+v9zHRx9f3UN6dXfzgbI9u5+ket/ZCtXHxseDWrWV14wjlsuOvOZ6JJddIPXNsbHItVp43PNUfOx8EFk7wsc1j+TwC3TgMzB+bg/VLQsfl2pk5PKBsRNU//y2d1DdJ4apzroEtEYOOR+zPYd6djN7yMz6zezFK27rMrNdZna4/n1lbDtCiOYyk7fx3wRw1+tu+wyAp9z9OgBP1X8XQixiomZ396cBXHrdzfcCeLj+88MAPjC/yxJCzDez/YCu293PAkD9e3AIm5ntMLPdZrZ7fIz/HyOEWDgW/NN4d9/p7r3u3tu+pHOhdyeECDBbs/eZ2ToAqH/nrVuFEE1ntmZ/HMD99Z/vB/D9+VmOEGKhiObZzewRAHcAWG1mpwB8DsCDAL5rZh8FcAIAH0ZdpzJZxfHj4d7v12zeROMPHHwxqC3vuZXGbr/nz6n+n//yBarnZJx3/+Hw7HYA6NrKe4wXLNJjPJIrr2bhXvyFSCm8FVhfdwCRmvNo7/csnCsvRXZdcH7tw4Wj/LgXySyAjTdfT2O3/P5fU/3QT/6V6oPlFVRf3rk1qG27mc8JyJ189kX6MkTN7u73BaT3xmKFEIsHXS4rRCLI7EIkgswuRCLI7EIkgswuRCI0tMS1UMjRumQoqB/Yy9P1tfFwKqWvyi/Fvf3m91H91t/7A6o/s2tXUBs/9xyNHV1xO9XbV4bbVANAqcRfk0skN5dFUmPIYrm5SCtqUlIJ8DbYsWnPk+dPUb0YG+lcCLfwHhrkrZ6v6eKjqoecl8jWavz5OD4RTkG/fJGPLu9qaw1qGUk36swuRCLI7EIkgswuRCLI7EIkgswuRCLI7EIkgswuRCI0NM8+WangxLGjQX3TmjU0/kjfsaB2y7XBzlgAgIPP8Rx+2TZS/frb3xbUDv0sPEoaAMZefoHqnbe9m+qxdLKTXHq1ystESwXertljT5FIO+gSHScdLs0FgIuHD1A9d77v8oZw6/Ge99xJYwcGwnlwAHj1xG6qD4+2UX1o8H+C2uoRXq7d3hFu5jw6MRrUdGYXIhFkdiESQWYXIhFkdiESQWYXIhFkdiESQWYXIhEammfPsxomBy4H9cOXeV61c/kNQe2l4z+nsW3l1VSvVEaovn5tOI+fFyO9njNe23z6AM/Zdr/xTVQvWvhhbCny2mg+PBhwi7WS5hcB1Ei9fAH8mGeRPHot0mN77fLwdRurT/JrHyY3raN6VuXXCEwYH+PdWQw/Zp3g+y6Tx6yA8DHRmV2IRJDZhUgEmV2IRJDZhUgEmV2IRJDZhUgEmV2IRGhonh1eQJ6Fe3lPjPKcbUdbOLc5NtxFYwtlflcvD4fHQQNAcSKcb97SeweNPfaL/6A6Bs5SuTrwZqoXyV13kFnTAPI80pM+0rPeCjzXXSUjhAcPHKSxsUL+ro1vofqJykBQKw720di1Z/g86UqNH5eLg/up/oa1Vwe1djtJY0eGw8fFM5aDj2BmD5lZv5m9eMVtD5jZaTPbU/+6J7YdIURzmcnb+G8CuGua27/s7tvrX0/M77KEEPNN1Ozu/jSASw1YixBiAZnLB3SfMLO99bf5waZYZrbDzHab2e5ajfdDE0IsHLM1+1cBXAtgO4CzAL4Y+kN33+nuve7eWyqFB9IJIRaWWZnd3fvcPXP3HMDXANw2v8sSQsw3szK7mV1Zg/dBADxvJYRoOtE8u5k9AuAOAKvN7BSAzwG4w8y2Y6qY+TiAj81kZ1lWwdDlcO/37quuofGXz4VjO9duprF9/a9SvdyxgurVQrg+edm6LTR26zX8fr1yhK/t1IGfUH3z298fFiN59GKkFr+WR+avR3Lh7aT2un/gAo3Njdfi23Je775p/R1BbYvzXPbhIb7vDR28nt3B9R5y3A8ceonGti9bGtSyavh4R83u7vdNc/M3YnFCiMWFLpcVIhFkdiESQWYXIhFkdiESQWYXIhEaWuJqZiiRUtP+ky/T+NZi+Aq8yng/jc0i3Z5b8nGqD14Mj+D1a3g55OWtfATv0ounqJ4P8LVd3PO/QW3Vtl6+7UhqLVLBitz5+eLc4fDazPnTj41cBoD+i7xF99bioaDmN2yjsZNHeKvpgSyc/gKA1svnqf70+XDa0cqdNLZlIuyDam0yqOnMLkQiyOxCJILMLkQiyOxCJILMLkQiyOxCJILMLkQiNDTPXjDDUpJnPz8yRuOLK7uDWjbJc661cd4SazLj+660h3PpB0ieGwBWbHwH1a9+1x9Rff/j36Z6dSSc0z35y300dsO2m6lejJwO8toA1ScvhFtZZ5GRy13tK6i+YiMvHcbkUFA6N8KfDzde81aqH9r7JNVbLaP6KCnfXQN+XcVAlbSS9nDLc53ZhUgEmV2IRJDZhUgEmV2IRJDZhUgEmV2IRJDZhUiEhubZa7UMfRcGg3prS5nG52QEb6HCX7dqzvOqY/lqqvd0hGuMrZ23HW7PR6neN8Bzsj1bV1H99MsXg1pxlLdMLmQ3Ud0jufDKCb79nLWaLi6nseeH+LUPWwv8uPRfPhzU1neEr9kAgMMXXqF6e3u4vwEAnOzj4xFrefi+DXXxtY1dOh3U8tocRjYLIX43kNmFSASZXYhEkNmFSASZXYhEkNmFSASZXYhEaGievVRqweo1G4L6+X4+qtYuh8fgrrvqahpbPM97cQ9OnuP6WDivWiS9ugGgYPwwD43xPHvvG99H9ZHLPwpqg+fD1zUAwMkXfkz1q3vvpPrQWT5ump1Pem6KjOge5j3t+0+H8+gAMDEZzjmPToTr7AGgWBmh+tFXea//1jI/j7a3hPPsI1X+fLEWkuMn10VEz+xmttHMfmxmB81sv5l9sn57l5ntMrPD9e8rY9sSQjSPmbyNrwH4tLvfCOBtAD5uZjcB+AyAp9z9OgBP1X8XQixSomZ397Pu/nz952EABwFsAHAvgIfrf/YwgA8s0BqFEPPAb/UBnZltBnALgGcAdLv7WWDqBQHA2kDMDjPbbWa7a+S6XSHEwjJjs5vZUgCPAviUu4c7+b0Od9/p7r3u3lsq8QGIQoiFY0ZmN7MWTBn9W+7+vfrNfWa2rq6vA8DHqAohmko09WZmBuAbAA66+5eukB4HcD+AB+vfvx/bVjWbRP+l40E9z3k55ZprbwhqAxcv09gs8i/E8k6emluxLFyOeWGYv9GZyPlhbjUe37XhvVRf95Zwem141w9obJm0JQaAc3ufprpFzhfjxfC7uexS5DHLeVlyLZLyhIXbKvcd5S22c0T2XeXlt9VwlhgAH4Vdbg2PZAYAq4RTtU48NJM8++0A/gTAPjPbU7/ts5gy+XfN7KMATgD40Ay2JYRoElGzu/tPAYReLvgpRwixaNDlskIkgswuRCLI7EIkgswuRCLI7EIkgjlr9TvPdHQs8xtv6g3qh0+cpfE3rA7nwtu6r6Wxx448R3Vv57nNpZ3hNtfr1m+nsRMX+qheXbON6uXqMaqvWrY0qC2d5LnoPU/9G9Vj54M8om960zuD2t4zh2hsi/NkdYmMPZ7SSVvlMm8FXavx50O5xI/rsgJf2wnSqrpj7fU0tlgLl9++evwQJsbHps2e6cwuRCLI7EIkgswuRCLI7EIkgswuRCLI7EIkgswuRCI0tJV0lucYGRsP6svaeG5ytBau4x0f4PueqPLXtSUrOqherhWDWnV8mMbWIqOFuwq8dvri6ADVT02E99+KNTS23Mm7B02MhGvCASAv8ON27nL4+oaO1o00tjrG691LrTwPPzEefr6Uiu00tmslv1/nzx2netty3mzZi+Hn+vggv96kiPBzMc/Cj5fO7EIkgswuRCLI7EIkgswuRCLI7EIkgswuRCLI7EIkQmPz7FkVgySH2FJeQeOXdIfHPQ+d209ju1d1Ub3awvPNHas2B7XBS0dpbKnEa6NPHjtB9byF9xzo2XBrUPOhF2hs69Z7qH5m36NUX7PlDVQ/cjRcs756A8+TV7NRqg9HRjp3Lp92IhkAYGiUj+hev4Zfn5BPhK8XAYDqkgs8vkauX+CXNmDJ0rBtC6Qhvc7sQiSCzC5EIsjsQiSCzC5EIsjsQiSCzC5EIsjsQiRCtG+8mW0E8E8AejCVAdzp7l8xswcA/BmA8/U//ay7P8G21Vpu856eTUE9s3D9MQB0r14X1E6ePk1jV3aFe6sDQKGth+qDl8K938vGe4iP5zxxunZZ+JgAwFiN13W3LlkW3vcAj7Uiz3V3dvLjNjgY7mEOAC2tpNd/G7/+oDLaT/WJMT4jPfNwr3+r8Ty5lXk9e16ZoDpa+HOiWgnnytuK/Bxs7eFa+XOnX0KlMn3f+JlcVFMD8Gl3f97MOgE8Z2a76tqX3f3vZrANIUSTmcl89rMAztZ/HjazgwDCl7IJIRYlv9X/7Ga2GcAtAJ6p3/QJM9trZg+Z2bTvLcxsh5ntNrPdWc7fpgshFo4Zm93MlgJ4FMCn3H0IwFcBXAtgO6bO/F+cLs7dd7p7r7v3Fgvh3llCiIVlRmY3sxZMGf1b7v49AHD3PnfP3D0H8DUAty3cMoUQcyVqdjMzAN8AcNDdv3TF7Vd+NP5BAC/O//KEEPPFTFJv7wTw3wD24f+L7z4L4D5MvYV3AMcBfKz+YV6Qlpayr1rdHdSLpXAKCQCWFMMtl2u2nMaOVXm7544y/6wyPPwXqPJO0FiykrfILkc2UCOtgwHAPFyeW17CS3fHJ3jq7eJlrmceLqkEAOQkBWV8bcVOfr/Xr+GfE6+shZ+Ox/t5iWuNtA4HgKUWabEd+eh7bDL8+VV7JCXpLWGfnDt5CJMTs0y9uftPAUwXTHPqQojFha6gEyIRZHYhEkFmFyIRZHYhEkFmFyIRZHYhEqGhraQBR56F87Z5zlsHD2ThssK2pW00dvWqFVQfOn+R6h1d4bbE4yVe7thW4nn2PFLa22J8vPDISPgqgLNnIqWY4Llu99hThOebrUByxsZz9EVy/QAA9J/kLbhPefhcdv0Nt9DY9rFXqH7mIr/flQq7MgOoZUNBrUrWDQBrulYEtQtnSJtpulUhxO8MMrsQiSCzC5EIMrsQiSCzC5EIMrsQiSCzC5EI0Xr2ed2Z2XkAr15x02oAfLZt81isa1us6wK0ttkyn2u72t2nnTfdULP/xs7Ndrt7b9MWQFisa1us6wK0ttnSqLXpbbwQiSCzC5EIzTb7zibvn7FY17ZY1wVobbOlIWtr6v/sQojG0ewzuxCiQcjsQiRCU8xuZneZ2UtmdsTMPtOMNYQws+Nmts/M9pjZ7iav5SEz6zezF6+4rcvMdpnZ4fr38Pzexq/tATM7XT92e8zsniatbaOZ/djMDprZfjP7ZP32ph47sq6GHLeG/89uZkUALwN4P4BTAJ4FcJ+7H2joQgKY2XEAve7e9AswzOxdAEYA/JO7v7F+2xcAXHL3B+svlCvd/W8WydoeADDS7DHe9WlF664cMw7gAwD+FE08dmRdf4wGHLdmnNlvA3DE3Y+6+ySA7wC4twnrWPS4+9MALr3u5nsBPFz/+WFMPVkaTmBtiwJ3P+vuz9d/Hgbw2pjxph47sq6G0AyzbwBw8orfT2FxzXt3AE+a2XNmtqPZi5mG7tfGbNW/h/tlNYfoGO9G8rox44vm2M1m/PlcaYbZp2s8tpjyf7e7+1sA3A3g4/W3q2JmzGiMd6OYZsz4omC248/nSjPMfgrAxit+vwrAmSasY1rc/Uz9ez+Ax7D4RlH3vTZBt/69v8nr+RWLaYz3dGPGsQiOXTPHnzfD7M8CuM7MtphZGcCHATzehHX8BmbWUf/gBGbWAeBOLL5R1I8DuL/+8/0Avt/Etfwai2WMd2jMOJp87Jo+/tzdG/4F4B5MfSL/CoC/bcYaAuu6BsAv61/7m702AI9g6m1dFVPviD4KYBWApwAcrn/vWkRr+2dMjfbeiyljrWvS2t6JqX8N9wLYU/+6p9nHjqyrIcdNl8sKkQi6gk6IRJDZhUgEmV2IRJDZhUgEmV2IRJDZhUgEmV2IRPg/mlF1CGmTEEIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Visualize dataset[first 5 images] from first batch of train_generator\n",
    "import matplotlib.pyplot as plt\n",
    "for i in range (0,5):\n",
    "    image = x_batch[i]\n",
    "    plt.imshow(image.astype(np.float32))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32, 28, 28, 3)\n",
      "(32, 15)\n"
     ]
    }
   ],
   "source": [
    "for image_batch, labels_batch in validation_generator:\n",
    "  print(image_batch.shape)\n",
    "  print(labels_batch.shape)\n",
    "  break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "S2fXxJz1I_SV"
   },
   "source": [
    "### Define model (10 Marks)\n",
    "- Initialize a Sequential Model\n",
    "- Add Convolution, Maxpool, Dropout, Flatten & Dense layers according to your model architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Conv2D, Dropout, MaxPooling2D, Flatten, Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def myownmodel_10layers():\n",
    "    model = Sequential()\n",
    "    model.add(Conv2D(input_shape=(IMG_SIZE,IMG_SIZE,3),filters=64,kernel_size=(3,3),padding=\"same\", activation=\"relu\", name = 'conv1'))\n",
    "    model.add(Conv2D(filters=64,kernel_size=(3,3), activation=\"relu\", name = 'conv2'))\n",
    "    model.add(Conv2D(filters=128, kernel_size=(3,3), activation=\"relu\", name = 'conv3'))\n",
    "    model.add(MaxPooling2D(pool_size=(2,2),strides=(2,2), name = 'pool1'))\n",
    "    # # Apply Dropout with 0.2 probability \n",
    "    model.add(Dropout(rate=0.2))\n",
    "    model.add(Conv2D(filters=256, kernel_size=(3,3), activation=\"relu\", name = 'conv4'))\n",
    "    model.add(MaxPooling2D(pool_size=(2,2),strides=(2,2), name = 'pool2'))\n",
    "    # # Apply Dropout with 0.2 probability\n",
    "    model.add(Dropout(rate=0.2))\n",
    "    model.add(Flatten(name='flatten'))\n",
    "    model.add(Dense(128, activation='relu', name='fc1'))\n",
    "    model.add(Dense(number_of_classes, activation='softmax', name='output'))\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Model with above definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "model=myownmodel_10layers()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "lRs36YJgJvI0"
   },
   "source": [
    "### Compile the model (5 Marks)\n",
    "- Specify optimizer, loss & metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### chosen categorical crossentropy for multi class classification problem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jvgO3JPvDQpL"
   },
   "outputs": [],
   "source": [
    "model.compile(loss=\"categorical_crossentropy\", metrics=[\"accuracy\"], optimizer=\"adam\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "U_Q6ihfUJ6Hm"
   },
   "source": [
    "### Get model summary (3 Marks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv1 (Conv2D)               (None, 28, 28, 64)        1792      \n",
      "_________________________________________________________________\n",
      "conv2 (Conv2D)               (None, 26, 26, 64)        36928     \n",
      "_________________________________________________________________\n",
      "conv3 (Conv2D)               (None, 24, 24, 128)       73856     \n",
      "_________________________________________________________________\n",
      "pool1 (MaxPooling2D)         (None, 12, 12, 128)       0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 12, 12, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv4 (Conv2D)               (None, 10, 10, 256)       295168    \n",
      "_________________________________________________________________\n",
      "pool2 (MaxPooling2D)         (None, 5, 5, 256)         0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 5, 5, 256)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 6400)              0         \n",
      "_________________________________________________________________\n",
      "fc1 (Dense)                  (None, 128)               819328    \n",
      "_________________________________________________________________\n",
      "output (Dense)               (None, 15)                1935      \n",
      "=================================================================\n",
      "Total params: 1,229,007\n",
      "Trainable params: 1,229,007\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "4mK8GNxFJ9vB"
   },
   "source": [
    "### Fit the model (5 Marks)\n",
    "- Specify epochs\n",
    "- Specify batch_size\n",
    "- Give validation_data\n",
    "- Validation accuracy should be more than 90%"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### calculate number of iterations per epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "508"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# steps_per_epoch = int(total train images / 32)\n",
    "steps_per_epoch = len(train_generator)\n",
    "steps_per_epoch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "170"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# valsteps = int(5757 / 32) #total validation image =5412  \n",
    "val_steps = len(validation_generator)\n",
    "val_steps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use earlystopping\n",
    "earlyStoppingcallback = tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=2, min_delta=0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9SLjprPDDYOs"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:sample_weight modes were coerced from\n",
      "  ...\n",
      "    to  \n",
      "  ['...']\n",
      "WARNING:tensorflow:sample_weight modes were coerced from\n",
      "  ...\n",
      "    to  \n",
      "  ['...']\n",
      "Train for 508 steps, validate for 170 steps\n",
      "Epoch 1/10\n",
      "508/508 [==============================] - 107s 210ms/step - loss: 0.6564 - accuracy: 0.7931 - val_loss: 0.2090 - val_accuracy: 0.9392\n",
      "Epoch 2/10\n",
      "508/508 [==============================] - 111s 219ms/step - loss: 0.0782 - accuracy: 0.9761 - val_loss: 0.1970 - val_accuracy: 0.9396\n",
      "Epoch 3/10\n",
      "508/508 [==============================] - 115s 225ms/step - loss: 0.0533 - accuracy: 0.9835 - val_loss: 0.1541 - val_accuracy: 0.9575\n",
      "Epoch 4/10\n",
      "508/508 [==============================] - 114s 224ms/step - loss: 0.0331 - accuracy: 0.9897 - val_loss: 0.1842 - val_accuracy: 0.9534\n",
      "Epoch 5/10\n",
      "508/508 [==============================] - 113s 221ms/step - loss: 0.0295 - accuracy: 0.9903 - val_loss: 0.1070 - val_accuracy: 0.9706\n",
      "Epoch 6/10\n",
      "508/508 [==============================] - 123s 242ms/step - loss: 0.0247 - accuracy: 0.9924 - val_loss: 0.1609 - val_accuracy: 0.9586\n",
      "Epoch 7/10\n",
      "508/508 [==============================] - 127s 249ms/step - loss: 0.0168 - accuracy: 0.9951 - val_loss: 0.1451 - val_accuracy: 0.9651\n"
     ]
    }
   ],
   "source": [
    "trafficmodel = model.fit(\n",
    "        train_generator,\n",
    "        steps_per_epoch=steps_per_epoch,\n",
    "        epochs=10,\n",
    "        validation_data=validation_generator,\n",
    "        validation_steps=val_steps,\n",
    "        callbacks=[earlyStoppingcallback])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Set the path where you want to store the model and weights \n",
    "model.save('traffic_model.h5')\n",
    "model.save_weights('traffic_model_weights.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-21-e2265053a7e4>:1: Model.evaluate_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use Model.evaluate, which supports generators.\n",
      "WARNING:tensorflow:sample_weight modes were coerced from\n",
      "  ...\n",
      "    to  \n",
      "  ['...']\n",
      "Test Accuracy: 96.508\n"
     ]
    }
   ],
   "source": [
    "_, acc = model.evaluate_generator(validation_generator, steps=len(validation_generator), verbose=0)\n",
    "print('Test Accuracy: %.3f' % (acc * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Getting accuracy of around 96.5% on validation set"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "diNbnH_dKO8J"
   },
   "source": [
    "### Draw plots (5 Marks)\n",
    "- Plot training accuracy and validation accuracy with respect to epochs\n",
    "- Plot training loss and validation loss with respect to epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAA670lEQVR4nO3de3hV5Zn///cnIQHCMYEAIQGCiAKKoCIeR620VvGAOtbaqdZSlTpVK07nOzq239GOnfn666XtdEZHq5WqHap1Wq1oqTZB0XoGNBIOKpFjSDgmgYRTDvv+/bFWwiYEshOy2dnJ/bqufe11eta6V5R17/U8az2PzAznnHMuVimJDsA551xy8cThnHOuTTxxOOecaxNPHM4559rEE4dzzrk28cThnHOuTTxxONcKSU9J+kmM266V9OV4x+RcInnicM451yaeOJzrJiT1SHQMrmvwxOG6hLCK6P9IWippl6QnJQ2V9GdJ1ZIKJWVGbX+5pOWSqiQtlDQ+at3Jkj4Ky/0O6NXsWJdKKgrLvivppBhjvETSx5J2Stog6b5m688J91cVrv92uLy3pIckrZO0Q9Lb4bLzJZW28Hf4cjh9n6TfS/ofSTuBb0uaKum98Bjlkh6WlB5V/gRJBZIqJG2WdI+kYZJ2SxoUtd2pkrZKSovl3F3X4onDdSV/C3wFOA64DPgzcA8wmOD/9e8DSDoOeBaYDWQD84GXJaWHF9E/Ar8BsoD/DfdLWPYUYA7wXWAQ8EtgnqSeMcS3C/gWMBC4BPh7SVeE+x0ZxvtfYUyTgaKw3IPAqcBZYUz/BERi/JvMAH4fHnMu0ADcSfA3OROYBnwvjKEfUAi8CgwHjgUWmNkmYCFwTdR+rwOeM7O6GONwXYgnDteV/JeZbTazjcBfgQ/M7GMz2we8CJwcbvd14E9mVhBe+B4EehNcmM8A0oD/MLM6M/s9sCjqGDcDvzSzD8yswcyeBvaF5Q7LzBaaWbGZRcxsKUHyOi9c/U2g0MyeDY+73cyKJKUA3wHuMLON4THfDc8pFu+Z2R/DY+4xsyVm9r6Z1ZvZWoLE1xjDpcAmM3vIzPaaWbWZfRCue5ogWSApFfgGQXJ13ZAnDteVbI6a3tPCfN9wejiwrnGFmUWADUBuuG6jHdj757qo6VHAD8KqnipJVcCIsNxhSTpd0hthFc8O4BaCX/6E+/iihWKDCarKWloXiw3NYjhO0iuSNoXVV/8eQwwALwETJB1DcFe3w8w+bGdMLsl54nDdURlBAgBAkggumhuBciA3XNZoZNT0BuDfzGxg1CfDzJ6N4bi/BeYBI8xsAPAY0HicDcCYFspsA/YeYt0uICPqPFIJqrmiNe/++lHgU2CsmfUnqMprLQbMbC/wPMGd0fX43Ua35onDdUfPA5dImhY27v6AoLrpXeA9oB74vqQekq4CpkaVfQK4Jbx7kKQ+YaN3vxiO2w+oMLO9kqYCfxe1bi7wZUnXhMcdJGlyeDc0B/iZpOGSUiWdGbapfA70Co+fBvwIaK2tpR+wE6iRNA74+6h1rwDDJM2W1FNSP0mnR61/Bvg2cDnwPzGcr+uiPHG4bsfMPiOor/8vgl/0lwGXmVmtmdUCVxFcICsJ2kNeiCq7mKCd4+FwfUm4bSy+B/yrpGrgXwgSWON+1wPTCZJYBUHD+KRw9T8CxQRtLRXA/wekmNmOcJ+/Irhb2gUc8JRVC/6RIGFVEyTB30XFUE1QDXUZsAlYBXwpav07BI3yH4XtI66bkg/k5JyLlaTXgd+a2a8SHYtLHE8czrmYSDoNKCBoo6lOdDwucbyqyjnXKklPE7zjMduThvM7Duecc23idxzOOefapFt0ejZ48GDLz89PdBjOOZdUlixZss3Mmr8b1D0SR35+PosXL050GM45l1QkrWtpedyqqiTNkbRF0rJDrJek/5RUoqBH01Oi1l0k6bNw3d1Ry7PCnjtXhd+ZLe3bOedc/MSzjeMp4KLDrL8YGBt+ZhF0hdDYbcIj4foJwDckTQjL3E3QW+dYYEE475xz7iiKW+Iws7cI3nI9lBnAMxZ4HxgoKYege4cSM1sdvsX7XLhtY5mnw+mngSviErxzzrlDSmQbRy4H9txZGi5raXljfzlDzawcwMzKJQ051M4lzSK4k2HkyJEHra+rq6O0tJS9e/ceyTm4o6RXr17k5eWRlubjBjmXaIlMHGphmR1meZuY2ePA4wBTpkw5qHxpaSn9+vUjPz+fAztCdZ2NmbF9+3ZKS0sZPXp0osNxrttL5HscpQRdWTfKI+ju+lDLATaH1VmE31vae/C9e/cyaNAgTxpJQBKDBg3yu0PnOolEJo55wLfCp6vOIBgYppygB9CxkkaHw3heG27bWOaGcPoGgsFl2s2TRvLw/1bOdR5xq6qS9CxwPjBYUilwL8GQnJjZYwTjPE8n6JZ6NzAzXFcv6TbgNSAVmGNmy8PdPgA8L+lGYD3wtXjF75xznVr9PthXDft2wr6acLoaamvCZeH8pG/AoBbH52q3uCUOM/tGK+sNuPUQ6+YTJJbmy7cD0zokQOecO9oa6qG2ev9FvemCvzO84Fcf/DnU8khdDAcUjDg9eRKH61h9+/alpqYm0WE41/1EIlC3q9nFfmezi3qzX/0H/fIP19Xvie2YaRnQsx+k9w2+e/aDgSP3Tzct7w89o7ZJ77d/umdfSOsDKR3fIuGJw7VJfX09PXr4/zauk6mvhbrdULcn/A6na3cdvKxuN9TuPsT24XRtzf6LfW2Mvcinph98Ae87DAYde+gLe+PFPzpBpPeF1M79b6xzR3eU/Pjl5awo29mh+5wwvD/3XnbCIdffddddjBo1iu9973sA3HfffUjirbfeorKykrq6On7yk58wY8aMQ+6jUU1NDTNmzGix3DPPPMODDz6IJE466SR+85vfsHnzZm655RZWr14NwKOPPsrw4cO59NJLWbYs6CHmwQcfpKamhvvuu4/zzz+fs846i3feeYfLL7+c4447jp/85CfU1tYyaNAg5s6dy9ChQ6mpqeH2229n8eLFSOLee++lqqqKZcuW8fOf/xyAJ554gpUrV/Kzn/3siP6+LolEGtpxAW/cfk/wa/+A7VvYT6S+7XGlZUBa7+BXeVrv4JPeB3pnwsARLV/Uoz/Nf/X3aG24967DE0eCXHvttcyePbspcTz//PO8+uqr3HnnnfTv359t27ZxxhlncPnll7f6RFGvXr148cUXDyq3YsUK/u3f/o133nmHwYMHU1ERvMj//e9/n/POO48XX3yRhoYGampqqKysPOwxqqqqePPNNwGorKzk/fffRxK/+tWv+OlPf8pDDz3E/fffz4ABAyguLm7aLj09nZNOOomf/vSnpKWl8etf/5pf/vKXR/rnS34N9VCzCWo2BxfWSANYA1gkatr2T0fCddYQVJ0037ZpfaTlfUUiUeWj1x9qX9bCts2+D1hv0FDbLBGEF/yG2rb/fXr0Ci/mGVEX+Azo1R/6DTt4XXpGC4kgI1wevW3G/iThT+q1mycOOOydQbycfPLJbNmyhbKyMrZu3UpmZiY5OTnceeedvPXWW6SkpLBx40Y2b97MsGHDDrsvM+Oee+45qNzrr7/O1VdfzeDBgwHIysoC4PXXX+eZZ54BIDU1lQEDBrSaOL7+9a83TZeWlvL1r3+d8vJyamtrm17KKyws5LnnnmvaLjMz6IPyggsu4JVXXmH8+PHU1dUxceLENv61klDtbthRCjvWB99VG2DHhv3TOzcGF9yjTpCSCkoFpeyfTkkJ5pXabH30spSobZuVV0pwsc8YtP/CHMsFvMULfu9gv67T8sSRQFdffTW///3v2bRpE9deey1z585l69atLFmyhLS0NPLz82N66e1Q5cws5vcfevToQSQSaZpvftw+ffo0Td9+++38wz/8A5dffjkLFy7kvvvuAzjk8W666Sb+/d//nXHjxjFz5syY4unUzGB3RQtJYcP+6d3bDyyjVOg/HAaMgJFnBFUhA0YEv55T0lq+GKekBr+KD7qYN/uO5cLetN5/Zbsj54kjga699lpuvvlmtm3bxptvvsnzzz/PkCFDSEtL44033mDduha7wj/Ijh07Wiw3bdo0rrzySu68804GDRpERUUFWVlZTJs2jUcffZTZs2fT0NDArl27GDp0KFu2bGH79u307duXV155hYsuarlz4x07dpCbmwvA008/3bT8wgsv5OGHH+Y//uM/gKCqKjMzk9NPP50NGzbw0UcfsXTp0iP4ix0lDfVQXX5gIjhgujSojomWlhEkggF5MHxyOD0iKkHkdPoGT+di5f8nJ9AJJ5xAdXU1ubm55OTk8M1vfpPLLruMKVOmMHnyZMaNGxfTfg5V7oQTTuCHP/wh5513HqmpqZx88sk89dRT/OIXv2DWrFk8+eSTpKam8uijj3LmmWfyL//yL5x++umMHj36sMe+7777+NrXvkZubi5nnHEGa9asAeBHP/oRt956KyeeeCKpqance++9XHXVVQBcc801FBUVNVVfJVTtrkPcKZQG0zvLDq5GyhgUJIDs4+HYL0clhTwYMBIysvzXvOs2FLyH17VNmTLFmo8AuHLlSsaPH5+giLqfSy+9lDvvvJNp09r//mZM/83Mgmqi5ncIVev3L9vTrLd/pUL/3CAJDBxxcFIYkBfUxTvXzUhaYmZTmi/3Ow4XV1VVVUydOpVJkyYdUdJoEokEDctV6/c3PkcniMNVIw0cAcNPCRPEyP1VS16N5Fyb+L+WJFJcXMz1119/wLKePXvywQcfJCii1g0cOJDPP/+8bYXMoKEOGvYF/fE0fqrL4d/OC5ZHyxgcJIXscXDsV/bfNTQmiN6ZXo3kXAfyxJFEJk6cSFFRUaLD6BgtJYfo6QOGYFHwclVKGpw+C7KOgYGjwruGvODxTefcUeOJw8WPWdARW0uJob4WiERtHCaH1J7Bm7g9eu6fT00L7hi2r4QpP0nU2TjnQp443JExC7p7qN8H9XujkkNtMG3Nk0N6VHIIp3v0DPr58eok55KCJw7Xuujk0LzdoaXkkJoeJIPG/ns8OTjXpXjicAFPDs65GMV16FhJF0n6TFKJpLtbWJ8p6UVJSyV9KOnEcPnxkoqiPjslzQ7X3SdpY9S66fE8h3ipqqriv//7v9tcbvr06VRVVbXvoI0N0vtqgncddpZBxRrY+ilsKobNy2D7quBR15rNQQd1qWnBy2/98yBrDAyZADmTYOiEYHCYAXnQJzvofK5HT08aznUD8Rw6NhV4BPgKUAoskjTPzFZEbXYPUGRmV0oaF24/zcw+AyZH7Wcj8GJUuZ+b2YPxiv1oaEwcjb3jNmpoaCA19dAdvM2ff9DAiDEecAPsqTz4jejGO4eMPuFdQ6+w7SE96NuojVqL3zmX/OJZVTUVKDGz1QCSngNmANGJYwLw/wDM7FNJ+ZKGmtnmqG2mAV+YWWwdN7XHn+8OfnF3pGET4eIHDrn67rvv5osvvmDy5MmkpaXRt29fcnJyKCoqYsWKFVxxxRVs2LCBvXv3cscddzBr1iwA8vPzWbx4MTU1NVx88cWcc845vPvuu+Tm5vLSSy/Ru3cLj6Y21PPEE4/z+G//SG19hGPHjOE3Tz9FRv+BbN6ylVtuOnBsjrPOOqvFcTy+/e1vc+mll3L11VcD+0clXLhwIT/+8Y9jiv/VV1/lnnvuoaGhgcGDB1NQUMDxxx/Pu+++S3Z2NpFIhOOOO47333+/qVdf51znEs/EkQtsiJovBU5vts0nwFXA25KmAqOAPCA6cVwLPNus3G2SvgUsBn5gZgf1CS5pFjALYOTIkUdwGvHxwAMPsGzZMoqKili4cCGXXHIJy5Yta+qifM6cOWRlZbFnzx5OO+00/vZv/5ZBgwYdsI9Vq1bx7LPP8sQTT3DNNdfwhz/8geuuu+7gg9VWc9XFF3DzHXdDeh9+9KMf8eQzc7n99ttbHJtj+fLlLY7jcTgffvhhq/FHIhFuvvlm3nrrLUaPHk1FRQUpKSlcd911zJ07l9mzZ1NYWMikSZM8aTjXicUzcbRU2d28Y6wHgF9IKgKKgY+BpqG8JKUDlwP/HFXmUeD+cF/3Aw8B3znoQGaPA49D0FfVYSM9zJ3B0TJ16tSmiy7Af/7nf/Lii0Ht3IYNG1i1atVBiWP06NFMnjwZgFNPPZW1a9e2vPO9O1n2+Rp+9Pf/QlVVFTU1NXz1q18FWh6b45lnnmlxHI8jjX/r1q2ce+65Tds17vc73/kOM2bMYPbs2cyZM6drdL3uXBcWz8RRCoyIms8DyqI3MLOdwEwABQM5rAk/jS4GPoquuoqelvQE8EqHR54A0eNdLFy4kMLCQt577z0yMjI4//zzWxyXo2fP/UNVpqamsmfPnoN3bAb7dvLtO+/lj/NeYdKkSTz11FMsXLjwkLEcalyN6DE7zIza2v0ju8US/6H2O2LECIYOHcrrr7/OBx98wNy5cw8Zm3Mu8eL5VNUiYKyk0eGdw7XAvOgNJA0M1wHcBLwVJpNG36BZNZWknKjZK4FlHR75UdCvXz+qq6tbXLdjxw4yMzPJyMjg008/5f3332//ger2QKSe6l27ycnJoa6u7oALc+PYHBA0bO/cuZNp06bx/PPPs317MBhRY1VVfn4+S5YsAeCll16irq6uTfGfeeaZvPnmm03dsEdXgd10001cd911XHPNNd64nmR27K7j90tKefLtNRRtqKKuIdJ6IZfU4nbHYWb1km4DXgNSgTlmtlzSLeH6x4DxwDOSGggazW9sLC8pg+CJrO822/VPJU0mqKpa28L6pDBo0CDOPvtsTjzxRHr37s3QoUOb1l100UU89thjnHTSSRx//PGcccYZ7T/QviAP3//jH3P66aczatQoJk6c2JS0DjU2R0vjeNx8883MmDGDqVOnMm3atAPuMqIdKv7s7Gwef/xxrrrqKiKRCEOGDKGgoACAyy+/nJkzZ3o1VZKo2l3LX5Zv5k/F5bxTso36yP7a4F5pKUzKG8hp+Vmcmp/JKSMzGdA7LYHRuo7m43F0ddtWQaQBhsQ2KFSiLF68mDvvvJO//vWvh9ym2/w366Qqd9XylxWb+FPxJt4Nk0VeZm+mT8xh+sQccgb0Ysm6ShavrWTxugqWl+2kIWJIcNyQfkzJzww+o7LIy+wd87DGLnF8PI7uKFIfjHbXd0iiIzmsBx54gEcffdTbNjqhil21vLZ8E/OLy3n3i+00RIwRWb258W9Gc8nEHCbmDjggATQmEYDdtfUUbagKE0kl84rKmPvBegCG9OsZ3JGMCpLJhJz+9EiN6/vIrgP5HUcXc+utt/LOO+8EM5EGaKjljjtmM3PW3yc2sA7QVf+bdTbba/bx2vLNzC8u573VQbIYmZXB9Ik5XDIxhxNz+7frbqEhYny2qZol6ypYHN6ZbKwKHujISE9l8oiBTBmVyZT8LE4eOZB+vbx6K9H8jqMFh3rKJ5k98sgj+2eq1sOeKhh2YsLi6Sjd4QdOIm2r2dd0Z/H+6goaIkb+oAy+e+4xTJ+YwwnD25csoqWmiAnD+zNheH+uPzMfgPIde1i8tpIl6ypZtLaCh98oIWKQIjh+WP8wkQTJJHegj7vSFpGIsbVmHwN6p9ErrWMfOOm2iaNXr15s376dQYMGdbnkAQSP4e7dGXRC2I6uQzoTM2P79u306tUr0aF0KVur9/Hq8k3MX1rOB2u2EzEYPbgPt5wXJIsJOUeeLFqTM6A3l03qzWWThgNQs6+eovVVLF5XweK1lbzwUSm/eX9duG0vpuRnMWVUJqeOymR8Tn9SU7rgv90YNSaG0srdlFbuCT/7pzdW7qG2IcL/3Hg654zt2Bdqu21VVV1dHaWlpS2+H9ElNNQFQ61mZEF630RHc8R69epFXl4eaWlefXEktlTv5bVlm/hTcTkfrqkgYnBMdh8uCdsmxg3r16l+SNU3RPh0UzWL1+6v3tq0M/g327dnD04eOTBoJxkVVG/16dl1fgtHIsaW6ujEcGCCKKvaS22zR58H900nNzODvMze4SeDC8YNaffd2qGqqrpt4ujy3vlPKPi/cOcKGJCb6GhcAm3ZuZc/h8li0doKzGBMY7I4KYfjh3auZHE4ZsbGqj1NT28tWlvBZ5urMQuqwsbn9GPKqKDR/bT8LIYN6Lx3qQ0RY0v13qa7g9gSQ88DkkJuOD0isze5AzPond6xVVKeOLpb4nj6Mti1Db73XqIjcQmweede/lxczvziTSxaFySLsUP6Nj31dNzQvkmTLFqzc28dH6+vCu5K1lZStKGKPXVBL9C5A3s3tZFMGZXJcUP7HbXqrejEUFq5m9KKMClUBQmirGoPdQ0HXn+bJ4YDksTA3h2eGFrjjePdyb4aWPcenJH8T1K52G3asZf5xeXMLy5nyfpKzOC4oX25Y9pYLpmYw9ih/RIdYlz075XGecdlc95x2QDUNURYWb6TRWsrWbKugve+2M5LRUFvR/169uCUUZlBO0l+JpNHDCQjvX2XwYaIsXnn3mbVSLvZWLXnkIkhu1+QGE7KG8j0iTkHJIjcgb07vBE7XjxxdEVr3oJIHRz75URH4uKsrGoPf14WPA21ZF3QSfS4Yf2488vHMX3iMI4d0jWTxeGkpaZwUt5ATsobyI3njMbM2FCxJ2hwX1fJkrWVPFTwOQA9UsQJw/tz6qis8OXETIb0D6q3DpUYGquTyqr2HPDGPATvp3SFxNAar6rqil75B1j6O/inNcGgTK5L2Vi1hz8Xl/On4nI+Xl8FBMmisc1iTHbyPwwRbzt21/HR+uAN90VrK/lkQxX76oP2hBFZvRE6bGLYX420vzppeBdKDI28qqq7MIOSAhh9XpdLGrtr6ylcuYU3Pt1CaorI6pNOZkY6WX3Swu90Mvukk5WRzoDeaaR0oUc1Syt38+fioIG7aEMVAONz+vOPFx7H9Ik5HOPJok0GZKTxpXFD+NK4oFeF2voIy8t2sGRdJR+tryQtNYXLJuU0tS101cTQXp44uprtJcGLf2fPTnQkHWJffQNvfraVl5eWU7hiM3vqGhjcN5301BS276pt+pXYXIogM2N/IsnskxaVaNIPSjSZfdLo27NHp2ow3lCxmz8vK+dPxZv4JEwWJwzvz//56vFMn5jD6MEtdzLp2i69Rwonj8zk5JGZiQ4lKXji6GpWBb3NJnP7Rn1DhPdWb2deURmvLt9E9d56MjPSuOqUXC6fNJzT8rOa7ib21DZQsbuWippaKnbXUrmrlopdtVTuPvB77bbdfLS+ispdtQdVPzRKS1VTQjkwqaSTlZEWfEclnaw+6R3+C3RDxW7+FDZwLy3dAcCJuf35p4uOZ/qJOeR7snCdgCeOrqakAAYfB5mjEh1Jm0QixpL1lbz8SRnzi8vZVlNLv549uPCEYVw2KYezjx1MWgud4PVOTyU3vXfMLziZGdX76psSTMUBiaYuWB4moJXlO6ncVUvVnjoO1RTYOy01TDJpZPXpuT/BNCacA+5ugiq15uexfvv+ZFG8MUgWJ+UN4O6LxzH9xBxGDspo2x/TuTjzxNGV1O6Gte/AaTclOpKYmBnLy3Yy75MyXvmkjLIde+nZI4Uvjx/KZZOGc/7x2R3+i14S/Xul0b9XGqMGxfbrvSFi7NhTd+CdzK5atoffTXc6u+tYu20Xlbtqqd5Xf8j99evVoymh7K1r4NNNwdgok/IG8M8Xj2P6xBxGZHmycJ1XXBOHpIuAXxAM5PQrM3ug2fpMYA4wBtgLfMfMloXr1gLVQANQ39iyLykL+B2QTzCQ0zVmVhnP80gaa9+Ghn1w7LRER3JYJVuqmVdUxstLy1mzbRdpqeLcsdn800Xj+PKEofTtZN1GNDbEZ/WJ/WGD2voIVbuDpBIkmroDqtIak1BGeio/nD6eiycOIy/Tk4VLDnH7FyopFXiEYBS/UmCRpHlmtiJqs3uAIjO7UtK4cPvoq96XzGxbs13fDSwwswck3R3O3xWv80gqJYXQozeMOjvRkRxkQ8VuXl5axryiMj7dVI0EZx4ziO+eewwXnTiMgRld6wmw9B4pDOnfq+mdAOe6knj+tJsKlJjZagBJzwEzCIaIbTQB+H8AZvappHxJQ81s82H2OwM4P5x+GliIJ45ASQGM/htI6xwXqy079/LK0nJeXlrW9L7BKSMHcu9lE7hkYo5fVJ1LUvFMHLnAhqj5UuD0Ztt8AlwFvC1pKjAKyAM2E4wp/hdJBvzSzB4Pyww1s3IAMyuX1OLwdpJmAbMARo4c2TFn1Jlt/wIqVsPpie1mpHJXLa8u38S8ojLeX7MdM5iQ05+7LhrHpSd53b1zXUE8E0dLD8Q3fzblAeAXkoqAYuBjoLFV8WwzKwsTQ4GkT83srVgPHiaaxyF4c7ytwSedkgXB99ij/xhuzb56ClZs4uVPynnr863UR4xjBvfh+xeM5bJJOd2y2wvnurJ4Jo5SYETUfB5QFr2Bme0EZgIoePNqTfjBzMrC7y2SXiSo+noL2CwpJ7zbyAG2xPEckkdJIWQdE3yOgr11DSz8bAvzPiljwcot7KuPMHxAL248ZzSXTRreISPGOec6p3gmjkXAWEmjgY3AtcDfRW8gaSCw28xqgZuAt8xsp6Q+QIqZVYfTFwL/GhabB9xAcLdyA/BSHM8hOdTtDTo2POVb8T1MQ4S3S7bx8idl/GX5Zmr21TO4bzrXnjaCyyYN55SRmV2qmw/nXMviljjMrF7SbcBrBI/jzjGz5ZJuCdc/BowHnpHUQNBofmNYfCjwYviLtQfwWzN7NVz3APC8pBuB9cDX4nUOSWP9u1C/Jy5vizdEjA/XVPDy0jL+XFxO5e46+vfqwfSJw7h8Ui5nHJNFjxZezHPOdV1xfWDezOYD85steyxq+j1gbAvlVgOTDrHP7Rz4yK5bVQipPSH/nA7ZnZnxSekO5hWV8afiMjbv3EfvtFS+MiF4Me/c4wbTs4d39uZcd9W53rRy7VNSAPlnQ/qRPbH06aadvPxJGS9/Us76it2kp6Zw3vHZXD5pONPGD2n3gDfOua7FrwTJrnIdbPscTp3ZruJrt+0KksXSMj7fXENqijhrzCBuv+BYLjxhGAN6p3VwwM65ZOeJI9mVFAbfbWjfKN+xhz8tLWfeJ2VNPbCelp/J/TNO4OKJOQzu2zMekTrnughPHJ2QmWEGETMi4XfjvBF+R4LvPp/9hdT+I9necwS2c+/+MpGgjBHM1zdEeH9NBS9/UsaitRWYwcTcAfxw+nguOSmH4TH2Luucc544DuOxN7/gT0vLg4tv5MALeONFPHo+Egkv+uy/6FtLF//oMgY0n49RGvV83HMhf2w4mx/9++sxlTl2SF/u/PJxXDZpuA8E5JxrF08ch9EnPZXBfdNJkZBEiiBFIiUFhFDjfPhNs/nGMvu3O7CMdPA+JCGilqUE24jobYLv4ZWL6LtkL7mnXc5Php7YSixw3NB+jBvWz1/Mc84dEU8ch3H9mflcf2Z+osM4tL/MgZQ0vnTR1dDTx5x2zh0d/uZWMitZAKPO9KThnDuqPHEkqx0bYctyOPYriY7EOdfNeOJIVl+EveHGoZsR55w7HE8cyWpVAfTPhSHjEx2Jc66b8cSRjBrqYPXCYGxxf0LKOXeUeeJIRqWLYN9Ob99wziWEJ45ktKoAUnrAMeclOhLnXDfkiSMZlRTCiNOh14BER+Kc64Y8cSSb6s2waWnQvuGccwkQ18Qh6SJJn0kqkXR3C+szJb0oaamkDyWdGC4fIekNSSslLZd0R1SZ+yRtlFQUfqbH8xw6nabHcL19wzmXGHHrckRSKvAI8BWgFFgkaZ6ZrYja7B6gyMyulDQu3H4aUA/8wMw+ktQPWCKpIKrsz83swXjF3qmtKoC+Q2HYxERH4pzrpuJ5xzEVKDGz1WZWCzwHzGi2zQRgAYCZfQrkSxpqZuVm9lG4vBpYCeTGMdbkEGmAL14PXvrzx3CdcwkSz8SRC2yImi/l4Iv/J8BVAJKmAqOAvOgNJOUDJwMfRC2+LazemiMps6WDS5olabGkxVu3bj2iE+k0Ni6BvVX+trhzLqHimTha+kncfLSJB4BMSUXA7cDHBNVUwQ6kvsAfgNlmtjNc/CgwBpgMlAMPtXRwM3vczKaY2ZTs7OwjOI1OZFUBKAWOOT/RkTjnurF4dqteCoyIms8DyqI3CJPBTAAFg0SsCT9ISiNIGnPN7IWoMpsbpyU9AbwSp/g7n5JCyDsNMrISHYlzrhuL5x3HImCspNGS0oFrgXnRG0gaGK4DuAl4y8x2hknkSWClmf2sWZmcqNkrgWVxO4POZNc2KPvYq6mccwkXtzsOM6uXdBvwGpAKzDGz5ZJuCdc/BowHnpHUAKwAbgyLnw1cDxSH1VgA95jZfOCnkiYTVHutBb4br3PoVL54HTBPHM65hIvrCIDhhX5+s2WPRU2/B4xtodzbtNxGgpld38FhJodVBZAxGHImJzoS51w352+OJ4NIJHjx79hpkOL/yZxzieVXoWRQ/jHs3u7VVM65TsETRzIoWQAIxnj/VM65xPPEkQxWFUDuKdBnUKIjcc652BKHpD9IukSSJ5qjbXcFbFzs1VTOuU4j1kTwKPB3wCpJD4QdErqjYfUbYBHvDdc512nElDjMrNDMvgmcQvDuRIGkdyXNDN/wdvGyqhB6ZwZVVc451wnEXPUkaRDwbYI3vD8GfkGQSAriEpkLHsMtKYQxF0BKaqKjcc45IMYXACW9AIwDfgNcZmbl4arfSVocr+C6vc3FsGuLt2845zqVWN8cf9jMXm9phZlN6cB4XLSSwuDbH8N1znUisVZVjZc0sHEmHPL1e/EJyTVZVQjDToJ+QxMdiXPONYk1cdxsZlWNM2ZWCdwcl4hcYO8O2PABjPWnqZxznUusiSMl7OocaBpPPP0w27sjtXohWIM/huuc63RibeN4DXhe0mME3ZnfArwat6hc0L7Rc0AwcJNzznUisSaOuwjGvfh7gu7O/wL8Kl5BdXtmQfvGmPMhNa493zvnXJvF+gJgxMweNbOrzexvzeyXZtbQWjlJF0n6TFKJpLtbWJ8p6UVJSyV9KOnE1spKypJUIGlV+J0Z68kmjS0robrMH8N1znVKsfZVNVbS7yWtkLS68dNKmVTgEeBiYALwDUkTmm12D1BkZicB3yJ4qbC1sncDC8xsLLAgnO9aSsJ3Kj1xOOc6oVgbx39N0F9VPfAl4BmClwEPZypQYmarzawWeA6Y0WybCQQXf8zsUyBf0tBWys4Ang6nnwauiPEckseqAhhyAvQfnuhInHPuILEmjt5mtgCQma0zs/uAC1opkwtsiJovDZdF+wS4CkDSVGAUkNdK2aGNb66H30NaOrikWZIWS1q8devWVkLtRPZVw/r3YazfbTjnOqdYE8fesEv1VZJuk3Qlh7hgR2lpzHBrNv8AkCmpCLidoA+s+hjLHpaZPW5mU8xsSnZ2dluKJtaatyBS59VUzrlOK9ZHdmYDGcD3gfsJqqtuaKVMKTAiaj4PKIvewMx2AjMBwvdE1oSfjMOU3Swpx8zKJeUAW2I8h+RQUgjpfWHEGYmOxDnnWtTqHUfYUH2NmdWYWamZzQyfrHq/laKLgLGSRktKB64F5jXb98BwHQS97r4VJpPDlZ3H/qR1A/BSDOeZHBofwx19HvTw9yudc51Tq4kjfOz21Og3x2NhZvXAbQQvD64Enjez5ZJukXRLuNl4YLmkTwmeoLrjcGXDMg8AX5G0CvhKON81bFsFO9Z7+4ZzrlOLtarqY+AlSf8L7GpcaGYvHK6Qmc0H5jdb9ljU9HvA2FjLhsu3A12zu1h/DNc5lwRiTRxZwHYOfJLKgMMmDtdGJYUw+HgYODLRkTjn3CHFlDjMbGa8A+n2anfD2ndgqnc67Jzr3GIdAfDXtPA4rJl9p8Mj6q7Wvg0N++DYrlkL55zrOmKtqnolaroXcCXNHq11R6ikANIyYORZiY7EOecOK9aqqj9Ez0t6FiiMS0Td1aoCyP8bSOuV6Eicc+6wYn1zvLmxgLfgdpTtX0DlGh/tzzmXFGJt46jmwDaOTQRjdLiOUBLevHn7hnMuCcRaVdUv3oF0ayWFkDUGso5JdCTOOdeqWMfjuFLSgKj5gZKuiFtU3UndXljzV3/pzzmXNGJt47jXzHY0zphZFXBvXCLqbta9A/V7vH3DOZc0Yk0cLW3ng2F3hJJC6NEL8s9JdCTOOReTWBPHYkk/kzRG0jGSfg4siWdg3UZJIYw6G9J6JzoS55yLSayJ43agFvgd8DywB7g1XkF1G5XrYNvnXk3lnEsqsT5VtQu4O86xdD/eG65zLgnF+lRVgaSBUfOZkl6LW1TdRckCGDgKBh2b6Eiccy5msVZVDQ6fpALAzCppfcxxdzj1+2D1m8HdRtvGyHLOuYSKNXFEJDV1MSIpnxZ6y21O0kWSPpNUIumgqi5JAyS9LOkTScslNY4/frykoqjPTkmzw3X3SdoYtW56jOfQuax/H+p2efuGcy7pxPpI7Q+BtyW9Gc6fC8w6XIFwrPJHCIZ3LQUWSZpnZiuiNrsVWGFml0nKBj6TNNfMPgMmR+1nI/BiVLmfm9mDMcbeOZUUQGp60LGhc84lkZjuOMzsVWAK8BnBk1U/IHiy6nCmAiVmttrMaoHngBnNdw30C8cz7wtUAPXNtpkGfGFm62KJNWmULICRZ0LPvomOxDnn2iTWxvGbgAUECeMHwG+A+1oplgtsiJovDZdFexgYTzC2RzFwh5lFmm1zLfBss2W3SVoqaY6kzEPEPEvSYkmLt27d2kqoR9mOUtiywp+mcs4lpVjbOO4ATgPWmdmXgJOB1q7GLbX4Nm8X+SpQBAwnqJp6WFL/ph1I6cDlwP9GlXkUGBNuXw481NLBzexxM5tiZlOys7NbCfUoK1kQfHv7hnMuCcWaOPaa2V4AST3N7FPg+FbKlAIjoubzOHjUwJnACxYoAdYA46LWXwx8ZGabGxeY2WYzawjvTJ4gqBJLLiUF0D8Psse1vq1zznUysSaO0vA9jj8CBZJeovWhYxcBYyWNDu8crgXmNdtmPUEbBpKGEiSj1VHrv0GzaipJOVGzVwLLYjyHzqGhLnwMd5o/huucS0qxvjl+ZTh5n6Q3gAHAq62UqZd0G/AakArMMbPlkm4J1z8G3A88JamYoGrrLjPbBiApg+CJrO822/VPJU0mqPZa28L6zm3Dh7Bvp1dTOeeSVpt7uDWzN1vfqmnb+cD8Zssei5ouAy48RNndwKAWll8fc7CdUUkBpPSA0eclOhLnnGuX9o457tqrpBBGnAG9+re+rXPOdUKeOI6m6k2wqdjHFnfOJTVPHEeTP4brnOsCPHEcTSUF0HcYDD0x0ZE451y7eeI4Whrq4Ys3vDdc51zS88RxtGxcAnurYKx3M+KcS26eOI6WkkJQChxzfqIjcc65I+KJ42gpKYC8qdC7xT4ZnXMuaXjiOBpqtkLZx94brnOuS/DEcTR88Xrw7e0bzrkuwBPH0VBSAH2yYdikREfinHNHzBNHvEUaghf/xkyDFP9zO+eSn1/J4q2sCPZUePuGc67L8MQRbyWFgGDMBYmOxDnnOoQnjngrKYDcU6DPQT3EO+dcUopr4pB0kaTPJJVIuruF9QMkvSzpE0nLJc2MWrdWUrGkIkmLo5ZnSSqQtCr87rwvRuyuCN4YP9Y7NXTOdR1xSxySUoFHCMYNnwB8Q9KEZpvdCqwws0nA+cBD4TCzjb5kZpPNbErUsruBBWY2FlgQzndOX7wOFvHecJ1zXUo87zimAiVmttrMaoHngBnNtjGgnyQBfYEKoL6V/c4Ang6nnwau6LCIO1rJAuidBcNPTnQkzjnXYeKZOHKBDVHzpeGyaA8D44EyoBi4w8wi4ToD/iJpiaRZUWWGmlk5QPg9pKWDS5olabGkxVu3bj3ys2mrSCRoGB9zAaSkHv3jO+dcnMQzcbTUd7g1m/8qUAQMByYDD0tqHFP1bDM7haCq61ZJ57bl4Gb2uJlNMbMp2dnZbQq8Q2wuhl1b/DFc51yXE8/EUQqMiJrPI7iziDYTeMECJcAaYByAmZWF31uAFwmqvgA2S8oBCL+3xO0MjsSqguDbh4l1znUx8Uwci4CxkkaHDd7XAvOabbMemAYgaShwPLBaUh9J/cLlfYALgWVhmXnADeH0DcBLcTyH9isphJxJ0LfFmjTnnEtaPeK1YzOrl3Qb8BqQCswxs+WSbgnXPwbcDzwlqZigausuM9sm6RjgxaDNnB7Ab83s1XDXDwDPS7qRIPF8LV7n0G57qmDDh3DOnYmOxDnnOlzcEgeAmc0H5jdb9ljUdBnB3UTzcquBFnsENLPthHcpndbqhWAN3r7hnOuS/M3xeCgphJ4DIO+0REfinHMdzhNHRzMLe8M9H1LjekPnnHMJ4Ymjo21ZAdVl3s2Ic67L8sTR0Zoew/X2Dedc1+SJo6OVFMLQE6F/TqIjcc65uPDE0ZH2VcP69/1uwznXpXni6Eir34RInScO51yX5omjI5UUQno/GHF6oiNxzrm48cTRUcyCxHHMedAjvfXtnXMuSXni6CjbPocdG7yayjnX5Xni6Cj+GK5zrpvwxNFRSgohexwMHNH6ts45l8Q8cXSE2l2w7h2/23DOdQueODrC2rehodYTh3OuW/DE0RFWFUBaBow6K9GROOdc3Hni6AglhTD6XOjRM9GROOdc3MU1cUi6SNJnkkok3d3C+gGSXpb0iaTlkmaGy0dIekPSynD5HVFl7pO0UVJR+Jkez3No1fYvoHKNV1M557qNuA0YISkVeAT4ClAKLJI0z8xWRG12K7DCzC6TlA18JmkuUA/8wMw+CsceXyKpIKrsz83swXjF3ib+GK5zrpuJ5x3HVKDEzFabWS3wHDCj2TYG9FMwuHhfoAKoN7NyM/sIwMyqgZVAbhxjbb+SQhh0LGSNTnQkzjl3VMQzceQCG6LmSzn44v8wMB4oA4qBO8wsEr2BpHzgZOCDqMW3SVoqaY6kzJYOLmmWpMWSFm/duvXIzuRQ6vbA2r/63YZzrluJZ+JQC8us2fxXgSJgODAZeFhS/6YdSH2BPwCzzWxnuPhRYEy4fTnwUEsHN7PHzWyKmU3Jzs5u/1kczrp3oH6vj/bnnOtW4pk4SoHo16jzCO4sos0EXrBACbAGGAcgKY0gacw1sxcaC5jZZjNrCO9MniCoEkuMVYXQoxfkn52wEJxz7miLZ+JYBIyVNFpSOnAtMK/ZNuuBaQCShgLHA6vDNo8ngZVm9rPoApKih9a7ElgWp/hbV1II+edAWu+EheCcc0db3J6qMrN6SbcBrwGpwBwzWy7plnD9Y8D9wFOSigmqtu4ys22SzgGuB4olFYW7vMfM5gM/lTSZoNprLfDdeJ3DYVWuhe2r4LSbEnJ455xLlLglDoDwQj+/2bLHoqbLgAtbKPc2LbeRYGbXd3CY7VNSGHx7w7hzrpvxN8fba1UhZObDoDGJjsQ5544qTxztUb8P1rwV3G2oxRsj55zrsjxxtMf696Bulz+G65zrljxxtMeqAkhNh9F/k+hInHPuqPPE0R4lC4Iu1NP7JDoS55w76jxxtNWOUti60p+mcs51W5442qrpMVxv33DOdU+eONpqVQH0z4Ps4xMdiXPOJYQnjrZoqIPVb8JYfwzXOdd9eeJoiw0fQG21V1M557o1TxxtUVIIKT2C8cWdc66b8sTRFqsKYeSZ0Kt/69s651wX5YkjVjvLYXMxHDst0ZE451xCeeKI1RcLgm9v33DOdXOeOGK1qgD65cDQExIdiXPOJVRcE4ekiyR9JqlE0t0trB8g6WVJn0haLmlma2UlZUkqkLQq/M6M5zkA0FAPq98Iqqn8MVznXDcXt8QhKRV4BLgYmAB8Q9KEZpvdCqwws0nA+cBDktJbKXs3sMDMxgILwvn42rgY9u7wbkacc4743nFMBUrMbLWZ1QLPATOabWNAv3CM8b5ABVDfStkZwNPh9NPAFXE8h0BJISgVjvlS3A/lnHOdXTwTRy6wIWq+NFwW7WFgPFAGFAN3mFmklbJDzawcIPwe0tLBJc2StFjS4q1btx7ZmawqgLzToPfAI9uPc851AfFMHC01Bliz+a8CRcBwYDLwsKT+MZY9LDN73MymmNmU7OzsthQ9UM1WKC8KuhlxzjkX18RRCoyIms8juLOINhN4wQIlwBpgXCtlN0vKAQi/t8Qh9v38MVznnDtAPBPHImCspNGS0oFrgXnNtlkPTAOQNBQ4HljdStl5wA3h9A3AS3E8h6B9o082DDsprodxzrlk0SNeOzazekm3Aa8BqcAcM1su6ZZw/WPA/cBTkooJqqfuMrNtAC2VDXf9APC8pBsJEs/X4nUORBqC0f6O+yqk+CsvzjkHcUwcAGY2H5jfbNljUdNlwIWxlg2Xbye8S4m7siLYU+GP4TrnXBT/GX04JQWAYMwFiY7EOec6DU8ch9N/OJz8TcjISnQkzjnXacS1qirpnfKt4OOcc66J33E455xrE08czjnn2sQTh3POuTbxxOGcc65NPHE455xrE08czjnn2sQTh3POuTbxxOGcc65NZNamYS6SkqStwLp2Fh8MbOvAcBLJz6Xz6SrnAX4undWRnMsoMztoQKNukTiOhKTFZjYl0XF0BD+XzqernAf4uXRW8TgXr6pyzjnXJp44nHPOtYknjtY9nugAOpCfS+fTVc4D/Fw6qw4/F2/jcM451yZ+x+Gcc65NPHE455xrE08chyHpIkmfSSqRdHei42kvSXMkbZG0LNGxHAlJIyS9IWmlpOWS7kh0TO0lqZekDyV9Ep7LjxMd05GQlCrpY0mvJDqWIyFpraRiSUWSFic6niMhaaCk30v6NPw3c2aH7dvbOFomKRX4HPgKUAosAr5hZisSGlg7SDoXqAGeMbMTEx1Pe0nKAXLM7CNJ/YAlwBVJ+t9EQB8zq5GUBrwN3GFm7yc4tHaR9A/AFKC/mV2a6HjaS9JaYIqZJf3Lf5KeBv5qZr+SlA5kmFlVR+zb7zgObSpQYmarzawWeA6YkeCY2sXM3gIqEh3HkTKzcjP7KJyuBlYCuYmNqn0sUBPOpoWfpPwVJykPuAT4VaJjcQFJ/YFzgScBzKy2o5IGeOI4nFxgQ9R8KUl6keqKJOUDJwMfJDiUdgurd4qALUCBmSXrufwH8E9AJMFxdAQD/iJpiaRZiQ7mCBwDbAV+HVYh/kpSn47auSeOQ1MLy5LyF2FXI6kv8AdgtpntTHQ87WVmDWY2GcgDpkpKumpESZcCW8xsSaJj6SBnm9kpwMXArWE1bzLqAZwCPGpmJwO7gA5rp/XEcWilwIio+TygLEGxuFDYHvAHYK6ZvZDoeDpCWIWwELgosZG0y9nA5WHbwHPABZL+J7EhtZ+ZlYXfW4AXCaqsk1EpUBp1F/t7gkTSITxxHNoiYKyk0WHD0rXAvATH1K2FDcpPAivN7GeJjudISMqWNDCc7g18Gfg0oUG1g5n9s5nlmVk+wb+R183sugSH1S6S+oQPXRBW61wIJOWTiGa2Cdgg6fhw0TSgwx4i6dFRO+pqzKxe0m3Aa0AqMMfMlic4rHaR9CxwPjBYUilwr5k9mdio2uVs4HqgOGwbALjHzOYnLqR2ywGeDp/eSwGeN7OkfpS1CxgKvBj8PqEH8FszezWxIR2R24G54Q/f1cDMjtqxP47rnHOuTbyqyjnnXJt44nDOOdcmnjicc861iScO55xzbeKJwznnXJt44nCuk5N0frL3Ouu6Fk8czjnn2sQTh3MdRNJ14RgbRZJ+GXZiWCPpIUkfSVogKTvcdrKk9yUtlfSipMxw+bGSCsNxOj6SNCbcfd+osRXmhm/RO5cQnjic6wCSxgNfJ+gkbzLQAHwT6AN8FHac9yZwb1jkGeAuMzsJKI5aPhd4xMwmAWcB5eHyk4HZwASCnk/PjvMpOXdI3uWIcx1jGnAqsCi8GehN0F16BPhduM3/AC9IGgAMNLM3w+VPA/8b9pOUa2YvApjZXoBwfx+aWWk4XwTkEwz+5NxR54nDuY4h4Gkz++cDFkr/t9l2h+vj53DVT/uiphvwf7sugbyqyrmOsQC4WtIQAElZkkYR/Bu7Otzm74C3zWwHUCnpb8Ll1wNvhmOLlEq6ItxHT0kZR/MknIuF/2pxrgOY2QpJPyIYPS4FqANuJRhA5wRJS4AdBO0gADcAj4WJIbrn0uuBX0r613AfXzuKp+FcTLx3XOfiSFKNmfVNdBzOdSSvqnLOOdcmfsfhnHOuTfyOwznnXJt44nDOOdcmnjicc861iScO55xzbeKJwznnXJv8//E43tgOg0tdAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# summarize history for accuracy\n",
    "plt.plot(trafficmodel.history['val_accuracy'],label = \"val_accuracy\")\n",
    "plt.plot(trafficmodel.history['accuracy'],label = \"train_accuracy\")\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['val_accuracy', 'train_accuracy'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## As you can see train accuracy in red colour is low at the starting but after 1 epoch it shoots to 98% and validation accuracy is also around 97% at the end of 4 epoch runs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# summarize history for loss\n",
    "plt.plot(trafficmodel.history['val_loss'])\n",
    "plt.plot(trafficmodel.history['loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['val_loss', 'train_loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## As you can see training loss is high at 0.61 when training starts but after 1 epoch it reduces to 0.05 and validation loss is also around 0.12 at the end of all epoch runs"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "Questions - Graded Project - German Traffic Sign.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}