Final Project/Code/Garbage_Classification_VGG16.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Garbage_Classification_VGG16.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/trong-khanh-1109/CS231.M13.KHCL/blob/main/Garbage_Classification_VGG16.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PCTCUUe7Ggg6",
        "outputId": "3b804631-578f-4d12-8987-41d4c529b3c3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1C8BEEW_Lw4AUj4tHYNEjKaJIA6nnqYS3\n",
            "To: /content/Garbage classification 1.zip\n",
            "100% 56.7M/56.7M [00:01<00:00, 37.4MB/s]\n"
          ]
        }
      ],
      "source": [
        "!gdown https://drive.google.com/uc?id=1C8BEEW_Lw4AUj4tHYNEjKaJIA6nnqYS3"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!unzip '/content/Garbage classification 1.zip'"
      ],
      "metadata": {
        "id": "9CakGyk3G8Z_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "metadata": {
        "id": "PuJrXkxXHwJO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Import library"
      ],
      "metadata": {
        "id": "S5YyH3H4JT8O"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import cv2\n",
        "from google.colab.patches import cv2_imshow\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.applications import VGG16\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Flatten, Dense, Dropout, GlobalAveragePooling2D, BatchNormalization\n",
        "from tensorflow.keras.optimizers import RMSprop, SGD, Adam, Nadam\n",
        "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint"
      ],
      "metadata": {
        "id": "YtQP7s40JRxd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "FJoin = os.path.join\n",
        "train_path = '/content/Garbage classification/train'\n",
        "validation_path = '/content/Garbage classification/validation'\n",
        "test_path = '/content/Garbage classification/test'\n",
        "path = '/content/Garbage classification'"
      ],
      "metadata": {
        "id": "-CqKOICCJYPo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Data preprocessing"
      ],
      "metadata": {
        "id": "yx2ebC4RJgiM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_datagen = ImageDataGenerator(\n",
        "        rescale = 1./255,\n",
        "        preprocessing_function=tf.keras.applications.vgg16.preprocess_input,\n",
        "        rotation_range = 20, # Dữ liệu được tạo ra sẽ được xoay ngẫu nhiên một góc trong phạm vi +rotation_rangeđến -rotation_range(tính bằng độ)\n",
        "        width_shift_range = 0.2, # Dịch theo chiều ngang ngẫu nhiên trong một phạm vi nào đó\n",
        "        height_shift_range = 0.2, # Dịch ảnh theo chiều dọc trong một phạm vi nào đó\n",
        "        horizontal_flip = True, # lật ảnh ngẫu nhiên theo chiều ngang\n",
        "        vertical_flip = True, # lật ảnh ngẫu nhiên theo chiều dọc\n",
        ")\n",
        "\n",
        "validation_datagen = ImageDataGenerator(\n",
        "        preprocessing_function=tf.keras.applications.vgg16.preprocess_input,\n",
        "        rescale = 1./255\n",
        ")\n",
        "\n",
        "test_datagen = ImageDataGenerator(\n",
        "        preprocessing_function=tf.keras.applications.vgg16.preprocess_input,\n",
        "        rescale = 1./255\n",
        ")"
      ],
      "metadata": {
        "id": "qyMKtwEYJbM3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img_shape = (224, 224, 3)\n",
        "train_batch_size = 16\n",
        "val_batch_size = 16\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "            path + '/train',\n",
        "            target_size = (img_shape[0], img_shape[1]),\n",
        "            batch_size = train_batch_size,\n",
        "            class_mode='categorical',\n",
        "            )\n",
        "\n",
        "validation_generator = validation_datagen.flow_from_directory(\n",
        "            path + '/validation',\n",
        "            target_size = (img_shape[0], img_shape[1]),\n",
        "            batch_size = val_batch_size,\n",
        "            class_mode='categorical',\n",
        "            shuffle=False)\n",
        "\n",
        "test_generator = test_datagen.flow_from_directory(\n",
        "            path + '/test',\n",
        "            target_size = (img_shape[0], img_shape[1]),\n",
        "            batch_size = val_batch_size,\n",
        "            class_mode='categorical',\n",
        "            shuffle=False,)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gExtYVEuKLNy",
        "outputId": "7809c8d8-dcf1-4729-c998-10c7f9277333"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 2880 images belonging to 6 classes.\n",
            "Found 960 images belonging to 6 classes.\n",
            "Found 961 images belonging to 6 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_generator.class_indices"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UFjit1A8MD33",
        "outputId": "8c106f69-cb70-439e-943d-5a3e8ec1be89"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'cardboard': 0, 'glass': 1, 'metal': 2, 'paper': 3, 'plastic': 4, 'trash': 5}"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Building model"
      ],
      "metadata": {
        "id": "chLmZVPjMBOP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_VGG16 = VGG16(weights = 'imagenet',\n",
        "              include_top = False,\n",
        "              input_shape = img_shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OP7wBWVHMI7v",
        "outputId": "d4083f42-f2f9-4cb6-c93a-5a41339aeedf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "58892288/58889256 [==============================] - 1s 0us/step\n",
            "58900480/58889256 [==============================] - 1s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_VGG16.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RqL3MqIxMUv6",
        "outputId": "4596daa4-267d-41a3-f3cd-ee571a76e1d3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"vgg16\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
            "                                                                 \n",
            " block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      \n",
            "                                                                 \n",
            " block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     \n",
            "                                                                 \n",
            " block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         \n",
            "                                                                 \n",
            " block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     \n",
            "                                                                 \n",
            " block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    \n",
            "                                                                 \n",
            " block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         \n",
            "                                                                 \n",
            " block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    \n",
            "                                                                 \n",
            " block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    \n",
            "                                                                 \n",
            " block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    \n",
            "                                                                 \n",
            " block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         \n",
            "                                                                 \n",
            " block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   \n",
            "                                                                 \n",
            " block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
            "                                                                 \n",
            " block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
            "                                                                 \n",
            " block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         \n",
            "                                                                 \n",
            " block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 14,714,688\n",
            "Trainable params: 14,714,688\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Fine tunning"
      ],
      "metadata": {
        "id": "jIbk8jULMzZ4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_VGG16.trainable = False"
      ],
      "metadata": {
        "id": "jfNtqcUBM2bq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model definition"
      ],
      "metadata": {
        "id": "4F6R2jcgM8Ey"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = Sequential()\n",
        "model.add(model_VGG16)\n",
        "\n",
        "# add global average pooling layer\n",
        "model.add(GlobalAveragePooling2D())\n",
        "\n",
        "model.add(Dense(units=512, activation='relu'))  # use ReLU activation function\n",
        "model.add(BatchNormalization())                 # normalize and scale inputs or activations\n",
        "model.add(Dropout(0.2))  \n",
        "\n",
        "model.add(Dense(units=6, activation='softmax'))\n",
        "model.summary()"
      ],
      "metadata": {
        "id": "CfpoSxYwM-kO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ac00725f-ebff-476c-cf35-1a5171859f48"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " vgg16 (Functional)          (None, 7, 7, 512)         14714688  \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 512)              0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " dense (Dense)               (None, 512)               262656    \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 512)              2048      \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 512)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 6)                 3078      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 14,982,470\n",
            "Trainable params: 266,758\n",
            "Non-trainable params: 14,715,712\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Compile model"
      ],
      "metadata": {
        "id": "5BnyTnoh1CFw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(optimizer=Adam(learning_rate=0.0001), loss='categorical_crossentropy', metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "UMvrVUTD1Bqy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Train the CNN model"
      ],
      "metadata": {
        "id": "UlAZ2KhY1J9e"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "History = model.fit(train_generator,\n",
        "          validation_data=validation_generator,\n",
        "          epochs=20,\n",
        "          verbose=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wI0EBntZ1MGU",
        "outputId": "c18dde75-7ffc-4950-c9d7-743210e9ba18"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "180/180 [==============================] - 71s 307ms/step - loss: 1.2166 - accuracy: 0.5569 - val_loss: 1.5189 - val_accuracy: 0.4375\n",
            "Epoch 2/20\n",
            "180/180 [==============================] - 56s 309ms/step - loss: 0.8598 - accuracy: 0.6962 - val_loss: 1.2341 - val_accuracy: 0.5448\n",
            "Epoch 3/20\n",
            "180/180 [==============================] - 56s 308ms/step - loss: 0.7631 - accuracy: 0.7337 - val_loss: 1.0678 - val_accuracy: 0.6187\n",
            "Epoch 4/20\n",
            "180/180 [==============================] - 54s 301ms/step - loss: 0.6853 - accuracy: 0.7622 - val_loss: 1.0703 - val_accuracy: 0.6187\n",
            "Epoch 5/20\n",
            "180/180 [==============================] - 54s 301ms/step - loss: 0.6617 - accuracy: 0.7646 - val_loss: 1.1408 - val_accuracy: 0.6167\n",
            "Epoch 6/20\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.6241 - accuracy: 0.7840 - val_loss: 0.9494 - val_accuracy: 0.6625\n",
            "Epoch 7/20\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.5778 - accuracy: 0.7944 - val_loss: 1.0251 - val_accuracy: 0.6469\n",
            "Epoch 8/20\n",
            "180/180 [==============================] - 54s 301ms/step - loss: 0.5863 - accuracy: 0.7920 - val_loss: 1.0018 - val_accuracy: 0.6573\n",
            "Epoch 9/20\n",
            "180/180 [==============================] - 54s 301ms/step - loss: 0.5565 - accuracy: 0.7944 - val_loss: 0.9907 - val_accuracy: 0.6562\n",
            "Epoch 10/20\n",
            "180/180 [==============================] - 54s 301ms/step - loss: 0.5407 - accuracy: 0.8031 - val_loss: 0.9198 - val_accuracy: 0.6771\n",
            "Epoch 11/20\n",
            "180/180 [==============================] - 54s 301ms/step - loss: 0.5350 - accuracy: 0.8101 - val_loss: 0.9272 - val_accuracy: 0.6854\n",
            "Epoch 12/20\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.5296 - accuracy: 0.8118 - val_loss: 0.8801 - val_accuracy: 0.6917\n",
            "Epoch 13/20\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.5110 - accuracy: 0.8177 - val_loss: 0.8459 - val_accuracy: 0.7031\n",
            "Epoch 14/20\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.4923 - accuracy: 0.8184 - val_loss: 0.9161 - val_accuracy: 0.6844\n",
            "Epoch 15/20\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.4880 - accuracy: 0.8309 - val_loss: 0.8639 - val_accuracy: 0.7063\n",
            "Epoch 16/20\n",
            "180/180 [==============================] - 54s 299ms/step - loss: 0.4698 - accuracy: 0.8316 - val_loss: 0.8702 - val_accuracy: 0.7000\n",
            "Epoch 17/20\n",
            "180/180 [==============================] - 54s 299ms/step - loss: 0.4812 - accuracy: 0.8344 - val_loss: 0.8264 - val_accuracy: 0.7240\n",
            "Epoch 18/20\n",
            "180/180 [==============================] - 54s 299ms/step - loss: 0.4771 - accuracy: 0.8382 - val_loss: 0.8543 - val_accuracy: 0.7083\n",
            "Epoch 19/20\n",
            "180/180 [==============================] - 54s 302ms/step - loss: 0.4328 - accuracy: 0.8538 - val_loss: 0.7783 - val_accuracy: 0.7417\n",
            "Epoch 20/20\n",
            "180/180 [==============================] - 54s 299ms/step - loss: 0.4521 - accuracy: 0.8451 - val_loss: 0.8599 - val_accuracy: 0.7042\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "loss = History.history['loss']\n",
        "validation_loss = History.history['val_loss']\n",
        "\n",
        "accuracy = History.history['accuracy']\n",
        "validation_accuracy = History.history['val_accuracy']"
      ],
      "metadata": {
        "id": "1GHr1oiD5tdF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Unfreeze the convolution base "
      ],
      "metadata": {
        "id": "BSf2__Ou6AXU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_VGG16.trainable=True"
      ],
      "metadata": {
        "id": "NOu1_E2I6Kx6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UDZXvkv36QBN",
        "outputId": "9bcb598c-d4fd-48af-ca72-5513418dd0eb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " vgg16 (Functional)          (None, 7, 7, 512)         14714688  \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 512)              0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " dense (Dense)               (None, 512)               262656    \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 512)              2048      \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 512)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 6)                 3078      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 14,982,470\n",
            "Trainable params: 14,981,446\n",
            "Non-trainable params: 1,024\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "History = model.fit(train_generator,\n",
        "          validation_data=validation_generator,\n",
        "          epochs=5,\n",
        "          verbose=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D651RJtc6Un3",
        "outputId": "81d6c564-5bf1-48f2-8f69-1878fe4fbfdb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "180/180 [==============================] - 55s 305ms/step - loss: 0.4430 - accuracy: 0.8427 - val_loss: 0.8255 - val_accuracy: 0.7281\n",
            "Epoch 2/5\n",
            "180/180 [==============================] - 55s 306ms/step - loss: 0.4335 - accuracy: 0.8441 - val_loss: 0.8147 - val_accuracy: 0.7323\n",
            "Epoch 3/5\n",
            "180/180 [==============================] - 56s 309ms/step - loss: 0.4260 - accuracy: 0.8490 - val_loss: 0.8456 - val_accuracy: 0.7177\n",
            "Epoch 4/5\n",
            "180/180 [==============================] - 54s 300ms/step - loss: 0.4308 - accuracy: 0.8517 - val_loss: 0.8892 - val_accuracy: 0.7104\n",
            "Epoch 5/5\n",
            "180/180 [==============================] - 54s 299ms/step - loss: 0.4037 - accuracy: 0.8601 - val_loss: 0.8259 - val_accuracy: 0.7292\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Visulization of Accuracy and Loss in Training and Validation sets"
      ],
      "metadata": {
        "id": "eB961sdT7e97"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "loss.extend(History.history['loss'])\n",
        "validation_loss.extend(History.history['val_loss'])\n",
        "\n",
        "accuracy.extend(History.history['accuracy'])\n",
        "validation_accuracy.extend(History.history['val_accuracy'])"
      ],
      "metadata": {
        "id": "u9AGrW2S7eMn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot the training and validation losses\n",
        "epochs = range(1, len(loss)+1)\n",
        "\n",
        "fig1 = plt.figure(figsize=(10,6))\n",
        "plt.plot(epochs,loss,c=\"red\",label=\"Training\")\n",
        "plt.plot(epochs,validation_loss,c=\"blue\",label=\"Validation\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.xticks(epochs)\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 422
        },
        "id": "0IsHu7847o5l",
        "outputId": "5e96393a-6af0-487e-ee41-f7832851d5ae"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f12d03ee690>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# plot the training and validation accuracy\n",
        "epochs = range(1, len(accuracy)+1)\n",
        "\n",
        "fig2 = plt.figure(figsize=(10,6))\n",
        "plt.plot(epochs,accuracy,c=\"red\",label=\"Training\")\n",
        "plt.plot(epochs,validation_accuracy,c=\"blue\",label=\"Validation\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.xticks(epochs)\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 422
        },
        "id": "x7DoNtvd7vV7",
        "outputId": "f234bee0-79e7-488b-b879-e31abb76012a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f1256cef2d0>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# save model \n",
        "model.save('/content/GarBage_Classification_VGG16_finetune.h5')"
      ],
      "metadata": {
        "id": "xaCJBOOo762d"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Test"
      ],
      "metadata": {
        "id": "pIcV18UQ75VW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import itertools\n",
        "def plot_confusion_matrix(cm, classes, normalize=False, title='Confusion matrix', cmap=plt.cm.Blues):\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1, keepdims = True)\n",
        "\n",
        "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    plt.title(title)\n",
        "    plt.colorbar()\n",
        "    tick_marks = np.arange(len(classes))\n",
        "    plt.xticks(tick_marks, classes, rotation=45)\n",
        "    plt.yticks(tick_marks, classes)\n",
        "\n",
        "    fmt = '.2f' if normalize else 'd'\n",
        "    thresh = cm.max() / 2.\n",
        "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "        plt.text(j, i, format(cm[i, j], fmt),\n",
        "                 horizontalalignment=\"center\",\n",
        "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
        "\n",
        "    plt.tight_layout()\n",
        "    plt.ylabel('True label')\n",
        "    plt.xlabel('Predicted label')"
      ],
      "metadata": {
        "id": "nBEtdQOg8Ztd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# load model\n",
        "model = keras.models.load_model(\"/content/gdrive/MyDrive/Học kỳ 5/Nhập môn thị giác máy tính/Garbage classification/Save_model/VGG16/GarBage_Classification_VGG16_finetune.h5\")"
      ],
      "metadata": {
        "id": "retskXbt8bFM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# accuracy \n",
        "score = model.evaluate(test_generator)\n",
        "print(\"Accuracy of test data: {:.2f}%\".format(score[1]*100))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IjXlUIDv8knx",
        "outputId": "d1d62e16-acda-481c-ed5d-04b64ba9775a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "61/61 [==============================] - 9s 149ms/step - loss: 1.3058 - accuracy: 0.6348\n",
            "Accuracy of test data: 63.48%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Plot Confunsion matrix"
      ],
      "metadata": {
        "id": "Gm40H56q8yOe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "Y_pred = model.predict(test_generator, 961 // val_batch_size+1)\n",
        "y_train_pred = np.argmax(Y_pred, axis=1)"
      ],
      "metadata": {
        "id": "VNhGlMhs82Fr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "labels = sorted(os.listdir('/content/Garbage classification/test'))\n",
        "\n",
        "print('Classification Report')\n",
        "print(classification_report(test_generator.classes, y_train_pred, target_names=labels, zero_division=1))\n",
        "\n",
        "print('Confusion Matrix')\n",
        "cnf_matrix = confusion_matrix(test_generator.classes, y_train_pred)\n",
        "\n",
        "plt.figure()\n",
        "plot_confusion_matrix(cnf_matrix, classes=labels, title='Confusion matrix of VGG-16')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        },
        "id": "Dp05dYFL86hK",
        "outputId": "8367a788-0305-413e-9db0-1a33d11c6fe5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification Report\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "   cardboard       0.60      0.91      0.73       160\n",
            "       glass       0.67      0.73      0.70       161\n",
            "       metal       0.68      0.57      0.62       159\n",
            "       paper       0.79      0.60      0.68       160\n",
            "     plastic       0.55      0.53      0.54       160\n",
            "       trash       0.56      0.47      0.51       161\n",
            "\n",
            "    accuracy                           0.63       961\n",
            "   macro avg       0.64      0.63      0.63       961\n",
            "weighted avg       0.64      0.63      0.63       961\n",
            "\n",
            "Confusion Matrix\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVYAAAEmCAYAAAA5jbhCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3gUVRfA4d9JQoDQkZ7QS+i9SBUFFZAuTZAuqKDYFRugfihi7wqKKAjYld6JKL33XoTQuyGUkOR8f8wElpBkl2Sr3pdnHnbuzM49m92czN65c6+oKoZhGIb7BPk6AMMwjH8bk1gNwzDczCRWwzAMNzOJ1TAMw81MYjUMw3Azk1gNwzDczCTWfzERySoi00TknIj8mIHj9BCRue6MzVdEpLGI7PDAcd3yszb+HUxi9QMi0l1EVovIeRE5IiKzRKSRGw7dCSgI3KKqndN7EFX9TlXvckM8HiUiKiJl0tpHVf9U1UgPVJ/qz1pEuonIfhGRZOUhInJcRFrb6zlE5F1731gROSAiP4lIPYfniIg8IiIbReSCiBwVkSgR6ZZaYCJSWESmishh+2dUIoV9movIWrveaBHpkrEfx3+bSaw+JiJPAu8Dr2P9YhYDPgXaueHwxYGdqhrvhmMFPBEJ8eDh0/pZ/wbkBm5LVt4CUGC2iGQGFgJVgNZATqACMAVo6fCcD4HHgaeAW4Bw4CX7WKlJBGYD96a0UUQqApOAF4FcQDVgTRrHM5xRVbP4aMH6EJ8HOqexT2asxHvYXt4HMtvbmgLRWL9kx4EjQF972ytAHHDFrqM/MAKY6HDsEli/2CH2eh9gLxAD7AN6OJT/5fC8BsAq4Jz9fwOHbVHAa8AS+zhzgXypvLak+J91iL890ArYCZwGXnDYvy6wDDhr7/sxEGpvW2y/llj79XZ1OP5zwFFgQlKZ/ZzSdh017fUiwAmgaSrxVrBf31lgC9A2tZ91Cs8dA4xLVvYD8J79+AH7NWVL47NQDkgAaqfz8xZi/4xKJCufBLzm69+Hf9Pi8wD+ywvWWUZ8UmJLZZ9XgeVAASA/sDTpl8BOEvH2PpnshHQByGNvH8H1iTT5egn7Fy0EyAb8A0Ta2woDlezHfbATK5AXOAP0tJ93n71+i709CthjJ4Gs9vqoVF5bUvzD7PgH2IltEpADqARcBEra+9cCbrXrLQFsAx53OJ4CZVI4/ptYf6Cy4pBY7X0GAFuBMGAO8HYqsWYCdgMvAKHAHVh/OCJT+tmm8PyG9s83q72ey35t1e31KcB4J5+Xh4D9Gfi8pZZY92L9MdyEldwnAnl9/fsRyItpCvCtW4CTmvZX9R7Aq6p6XFVPYJ0d9XTYfsXefkVVZ2KdMaW3DTERqCwiWVX1iKpuSWGfe4BdqjpBVeNVdTKwHWjjsM/XqrpTVS9inZVVT6POK8BIVb2ClVzyAR+oaoxd/1asr6ao6hpVXW7Xux/4ghu/Xqf0moar6mU7nuuo6lishLkC64/Ji6kc51YgO9YfiThVXQhMx/rD4pSqLgGOAR3soi5YTQfr7fV8WGfVAIhIdRE5KyL/OFxsu24fe79oe79LIlLclVhSEIH1mboXKIv1B+ijdB7LwLSx+topIJ+Ttr8iwN8O63/bZVePkSwxX8BKADdFVWOxvj4/BBwRkRkiUt6FeJJiCndYd/zldxbPKVVNsB8nJb5jDtsvJj1fRMqJyHT7gs0/WO3S+dI4NsAJVb3kZJ+xQGXgI1W9nMo+RYCDqproUJb8dTvzLdDLftzTXk9yCiuxA6Cq61U1N9AR62z7hn3s/SKwfgaZsa5tNbYvgp4XkZT+MKbkItf+GJ7H+rm2uonXZSRjEqtvLQMuY7UrpuYw1oWRJMXssvSIxfrKm6SQ40ZVnaOqd2L98m7HSjjO4kmK6VA6Y7oZn2HFVVZVc2J9LZe0n0Kaw7eJSHasduuvgBEikjeVXQ8DRUXE8XfmZl/3BKCZiNTHOgP+zmHbAuAuEcmWxvMXAhEiUju1HdTq9ZDdXiq5GNdGrv85mSHvMsgkVh9S1XNY7YufiEh7EQkTkUwi0lJERtu7TQZeEpH8IpLP3n9iOqtcDzQRkWIikgt4PmmDiBQUkXb2L/ZlrCaFxBSOMRMoZ3cRCxGRrkBFrK/FnpYDq53yvH02/XCy7ceAUjd5zA+A1ar6ADAD+DyV/VZgnX0/a79HTbGaP6a4WpHdfPEX1ns6T1Udz+y/xWrf/FVEKotIsIhkAWo7PH8HVvPHFBG50+47G4x1MTFN9rGSznwz2+tJvgb6ikgpEQkDhuKd9/NfyyRWH1PVd4AnsbrMnAAOAo9gddEB+B+wGuusYhOw1i5LT13zgO/tY63h+l+eIDuOw1hXym/jxsSFqp7C6g70FNZX02eB1qp6Mj0x3aSnge5YF43GYr0WRyOAb+w2R6f9MEWkHdYFxKTX+SRQU0R6JN9XVeOwEmlL4CRWl7heqrr9Jl/DN1hn/I7NANjNFbdjtSnPwPoDsgOog9Uem2QwVperd7Hep2isC09dgQNp1HsR648lWGf9V9ubVXWcHc8KrOaNy8CQm3xdhgNRNWf9hmEY7mTOWA3DMNzMJFbDMAw3M4nVMAzDzUxiNQzDcDNPDkrxrychWVVCc/g0hhoVivm0/iT+cAk00U8uxAaJs661/w0H/t7PyZMn3frDCM5ZXDX+hhvobqAXT8xR1bQGpvEok1gzQEJzkDnSt6OrLVnxsU/rT5KQ6PukdjEuwflOXpAtc7CvQ8AP3g4a16/j9mNq/EWXfucurf/E2R15HmUSq2EYgUMEgnz/h8sZk1gNwwgs4v+XhkxiNQwjsARAG7b/p37DMIyrxDpjdbY4O4rIOHtanM0pbHvKnsImn70uIvKhiOy2p8Sp6ez4JrEahhE4BKuN1dni3HhSmM5GRIoCd3H9uAstscapLQsMxBplLU0msRqGEUDEagpwtjihqouxBrFJ7j2sgYUc+1W0A75Vy3Igt4gUTuG5V5nEahhGYHFDU0CKh7VGOzukqhuSbQrHGnUuSTROBjg3F68Mwwgsrl28yiciqx3Wx6jqmNQPKWFYA6e7ZZp3c8bqQZ8P78HfC95g9Y8v3LDtsZ53cHHdx9yS+9qA8Y1rlWX5lKGs+elF5n75mFdinDtnNlUrRVKpfBneGj3KK3U6ij54kJZ33UGtapWoXb0yn3z0gddjAPji0w9pXLc6jepU4/NPfBPDgwP6UTy8ILWrV/FJ/eA/70eqkvqxOm9jPamqtR2WVJOqrTRQEtggIvux5gFbKyKFsGaJKOqwbwROZo4widWDJkxbTrvBn9xQHlEwN81urcCBI9eaeHJlz8oHL3Sh8+NfUKvTSHo885XH40tISODxIYP5fdos1m3cyo9TJrNt61aP1+soJCSEN958mzUbtrDoz2WM/fxTtm3zbgzbtm5m4vhxzIlaStSyNcybPZO9e3Z7NQaAnr368Nv0WV6v15E/vB9OeaApQFU3qWoBVS2hqiWwvu7XtGd5mAr0snsH3AqcU9UjaR3PJFYPWrJ2D6fPXbihfPTT9/LiB7/hOMh415a1+X3BBg4ePQPAiTPnb3ieu61auZLSpctQslQpQkND6dy1G9On/e7xeh0VKlyY6jWs3is5cuQgsnwFjhzyxvRZ1+zcsZ2atesQFhZGSEgIDRo1YcbU35w/0c0aNW5C3jypTbnlHf7wfqTNbd2tJmPNORdpz3TbP43dZ2JNEb4ba+aKQc6Ob9pYvax10yocPn6WTTuv/7CWLV6AkJBg5ox9jOxhmflkchSTpq/0aCyHDx8iIuLaN5zw8AhWrlzh0TrT8vf+/WzYsI7adet5td4KFSrx+ivDOH3qFFmyZmX+nFlUq1nLqzH4I1+9H2kSIDjjt7SqaprTlttnrUmPFWtKHJcFRGK12zxqY02DPF1VK3u4vhKeqCdrlkw82+9uWg+6ceCUkOAgalYoSssHPyJrlkxEffMUKzfuZ/eB4+4MwW+dP3+eHt068ebb75EzZ06v1l2ufAUefeJpOrdvSVhYNipXrUawG355A5kv3w+nzJ1XN09EvJ7svVVnqYj8FA+/hZXfP8/2Ga8QXiA3yyY9R8FbcnDo+FnmLdvGhUtxnDoby19rd1O13M1MWX/zihQJJzr6Wi+SQ4eiCQ/3bJ0puXLlCj26dqJrt+60a9/R6/UD3N+7Hwv+XMm0OYvIlTsPpcuU9Ukc/sAf3o/UuacpwNM8GoGI9LJvAdsgIhNEpI2IrBCRdSIyX0QK2vuNsLcvASaIyC0iMldEtojIl1w/d3yIiHwnIttE5Ce7mwQi0sw+7ib7drXMdvkwEVklIptFZIyI9edORKJE5H27S8ZjIlLLjnMDN3na76otuw9TvNnzlL9nOOXvGc6h42ep3/1Njp2KYVrURhpUL01wcBBZs2SiTuUSbN931PlBM6B2nTrs3r2L/fv2ERcXx4/fT+Ge1m09WmdyqsqgBx8gsnx5Hn38Sa/W7ejECeubQfTBA8yY+hv3dk7zm+K/lr+8H2lyww0CnuaxxCoilbCmdL5DVasBj2HNqX6rqtbAmo/9WYenVASa220fw4G/VLUS8CvgOJpzJPCpqlbAmiJ4kD1H+nigq6pWwWriSJrS+GNVrWN/rc+KNXVzklC7K8Y7WHOrP2rHmtbrGigiq0VktbMBd795ow9R3zxFueIF2T37NXq3r5/qvjv2HWPe0q2s+uF5/pzwDON/XcrWPWleeMywkJAQ3vvgY9rcczfVq1Tg3s5dqFipkkfrTG7Z0iVM/m4Cf0Qton6dGtSvU4M5s2Z6NQaAvj260LB2Ve7v0p433/2QXLlzez2G3vd3p2mTBuzcuYMyJYsy/mvP9wxJzl/ej1S53t3Kt2F6avprEXkUKKSqLzqUVQHeAQoDocA+VW0hIiOw2ohfsfdbD3RU1b32+mmgHFYb62JVLWaX34E1//lw4CNVbWKXNwMGq2pHEbkXK4GHAXnt/UaJSBQwXFX/EJHcwEaH41YFJjlrYw0KK6C+Huj6zCoz0HUSM9D1NX7wdtC4fh3Wrlnt1tPHoFxFNXP9J5zud2nOU2tUtbY7674Z3m6M+AjrDLIK8CCQxWFbrIvHSP6RSfUjZJ/Jfgp0suscm846DcPwF//lpgBgIdBZRG4BEJG8QC6u3bHQO43nLga6289rCeRx2FZMRJK+U3fHal7YAZQQkTJ2eU/gD64l0ZMikh3olFJlqnoWOCsijeyiHi69QsMwvCwwmgI8djVcVbeIyEjgDxFJANYBI4AfReQMVuItmcrTXwEmi8gWYCnXD+G1AxgsIuOArcBnqnpJRPraxw4BVgGfq+plERkLbAaO2uWp6QuMExEF5qbvVRuG4VGCX1z1d8aj3YxU9Rvgm2TFN9zao6ojkq2fIuXBEE4C5VOpawFQI4Xyl7AuoiUvb5psfQ3geOHqWQzD8DNiEqthGIbb+UEbqjMmsRqGEVj8oA3VGZNYDcMIHGKaAgzDMNzPNAUYhmG4l5jEahiG4T5WS4BJrIZhGG4k5ozVMAzD3UxiNQzDcLOgINMrwDAMw32E60dn9lMmsWZAjQrFWLLCt8P2FXvwB5/Wn2T5qDa+DoEiebL6OgQAjpy95OsQyBOWydch4IkRScVNbaz2WCOtgeNJw4OKyFtAGyAO2AP0tQdoQkSeB/oDCcAQVZ2T1vH9/5zaMAzDgYg4XVwwHmiRrGweUFlVqwI7geft+ioC3YBK9nM+FZE0b/8yidUwjIASFBTkdHFGVRcDp5OVzVXVeHt1ORBhP24HTFHVy6q6D2sa7LppxnizL8owDMNnxMUF8iVNoWQvA2+ypn7ALPtxOHDQYVu0XZYq08ZqGEZAcfGr/sn0Ts0iIi8C8cB36Xk+mMRqGEYAEcSj3a1EpA/WRa1mem1CwENAUYfdIrg2E0qKTFOAYRiBxbWmgJs/rEgLrAHu26rqBYdNU4FuIpJZREoCZYGVaR3LnLEahhE4xD13XonIZKApVltsNNZMz88DmYF5dh3LVfUhe5qpH7CmgorHmgE6zSmBTWI1DCOguCOxqup9KRR/lcb+I4GRrh7fJFbDMAKGp9tY3cX/I/yXmjtnNlUrRVKpfBneGj3Ko3W937cOW95ryx+v3n21rE3tCBa/ejdHx3amWvFrs4uHBAsf9atL1Ct38ddrLRjSKsW5GzPkuccepE7F4rRocu2i7bbNG+nUsiktb6vDgPvvJSbmH7fXmxZvvh+Onh3yIHUqFKNF41pXy7Zu2kDHFk24p2k92jZvyIa1aU0u7H7nzp6ld48u1K1RiXo1K7NyxTKv1u+Uh9pY3ckkVh9ISEjg8SGD+X3aLNZt3MqPUyazbetWj9U3Zck+ur23+Lqy7YfO0feTpSzbeeK68ra1ixKaKYimw+dy52vz6HVbaYreEubWeO7t1pOvp/x2XdnzTw7imZdfY9Yfq7irVVvGfvKeW+tMi7ffD0eduvXk6ynXT1w86tUXGfL0i8yIWsETz73MqFde9EosSYY+8wTN7rybleu28OfytURGVvBq/WkSt9155VEmsfrAqpUrKV26DCVLlSI0NJTOXbsxfdoNs4K7zfKdJzkbG3dd2a4jMew5FnPDvqpKWGgIwUFClkzBXIlPJOZS/A37ZUTd+o3InTvvdWX79uymbv1GADS8rRlzpnvu55Gct98PR3UbNCJ3nut/FoJw3j5jj4k5R4FChb0SC8C5c+dYuuRPevbuB0BoaCi5cuf2Wv2uMInVSNHhw4eIiLjWLS48PIJDh9LsFuc109ZEcyEunk3vtmHtW635dM6OG5KyJ5SNrMC8WdMAmDX1F44civZ4nUn87f14eeRbvPHKCzSsVoY3hj/Psy+96rW6D+zfR758+Rj8YH+a1K/NkEEDiY2N9Vr9rpAgcbr42r8ysYrIeBHp5Os4AlHNknlJTFSqPjWNOs/N4OG7y1E8XzaP1/vmB5/z3ddjadu8AbHnY8gUGurxOv3Vd1+P4aXXRrNkw25eem00zz3+sNfqjk+IZ8P6dfQb8CCLl60mLCwb77/zptfqd4U5YzVSVKRIONHR1249PnQomvDwNG899pqO9YqzcPNR4hOUkzGXWbn7FNVK5HH+xAwqXTaSb36cxtT5S2nTsQvFSpT0eJ1J/O39+Pn772jRuj0Ardrdy8a1q71Wd5EiERQJj6B2nXoAtO3QkQ3r13mtfmdcSaomsbqBiLwsIjtE5C8RmSwiTyfbPkxEVonIZhEZI/ZPXUSGiMhWEdkoIlPssttEZL29rBORHJ6IuXadOuzevYv9+/YRFxfHj99P4Z7WbT1R1U07dPoCjcoXACAsNJhapfKy++iNbbHudvLEcQASExP5+N036d77AY/XmcTf3o+ChQqzYumfACz9M4oSpcp4se5ChEdEsGvnDgAWRy0ksrwfXbzCPaNbeVpA92MVkTrAvUA1IBOwFliTbLePVfVVe/8JWPcBTwOGAiVV9bKIJLXOP411V8USEckOeGTE4pCQEN774GPa3HM3CQkJ9O7Tj4qVKnmiKgA+H3grDSPzkzd7Zta/1ZrRv2/hbGwcr3evwS05MjPpscZsPniWru8tZtzC3XzQrw6LX70bEZjy1362Rp9zazyPPdibFUsWc+b0KRpWK8Njz75EbGwsE8d9AcDd97Sj03293FpnWrz9fjgaMrAXK5b8yZnTJ2lQtTSPPfsyr7/7Ca+9+AzxCfFkzpyZke96dzD10W9/wMB+vYiLi6NEyZJ88nmq/eZ9w/cnpE6JemKYby8RkceBPKo63F5/FzgMVAamq+pPInIv1v2/YUBe4CNVHSUis4HzwG/Ab6p6XkSGAh2wRrX5RVVvuIJiDz82EKBosWK1du752+OvMy1mBoFrzAwC1/jDDAK3N6rHurWr3ZoGMxcsq+E9PnC637737lmT3tGt3MH358weJCJZgE+BTqpaBRgLZLE33wN8AtQEVolIiKqOAh4AsgJLROSG3vGqOkZVa6tq7fz58nvldRiGYTP9WL1iCdBGRLLYX91bJ9uelERP2ts7AYhIEFBUVRcBzwG5gOwiUlpVN6nqm8AqwP23HRmGkW7WLa3OF18L6DZWVV0lIlOBjcAxYBNwzmH7WREZC2wGjmIlS4BgYKKI5MJqsfnQ3vc1EbkdSAS2cG0EccMw/IQfnJA6FdCJ1fa2qo4QkTBgMbBGVccmbVTVl4CXUnheo+QFqvqo58I0DMMd/OGrvjP/hsQ6xp5FMQvwjaqu9XVAhmF4hggEB5vE6nGq2t3XMRiG4T0BcMIa+InVMIz/FtMUYBiG4U4SGGesgd7dyjCM/5CkGQQyekuriIwTkeMistmhLK+IzBORXfb/eexyEZEPRWS3fQt8TWfHN4nVMIyAIuJ8ccF4oEWysqHAAlUtCyyw1wFaYs3MWhbrrsvPnB3cJFbDMAKKO+68UtXFwOlkxe2Ab+zH3wDtHcq/VctyILeIpDn6uEmshmEEDhfOVjPQBltQVY/Yj48CBe3H4cBBh/2i7bJUmYtXhmEEDAFXb1nNJyKOA9mOUdUxrtajqioi6R6hyiRWwzACiovdrU6mY3SrYyJSWFWP2F/1j9vlh4CiDvtF2GWpMk0BhmEEFA82BUwFetuPewO/O5T3snsH3Aqcc2gySJE5Y82Ay/GJHDh5wacxzB/R0qf1JxkwZb2vQ+DXAfV8HQIAebP5fizUKwm+H2dZcX8MIi43BTg5jkwGmmI1GUQDw4FRwA8i0h/4G+hi7z4TaAXsBi4AfZ0d3yRWwzACiHvGW1XV+1LZ1CyFfRUYfDPHN4nVMIyAEgh3XpnEahhGQDFjBRiGYbiRu9pYPc0kVsMwAoo5YzUMw3CzAMirJrEahhFATFOAYRiGe4mbult5mkmshmEElADIqyaxesuRQ9E8O2QAp04cR0Tocn9feg8YzJuvvsCiubPIFJqJYsVL8cb7n5MzV26PxHD0cDQvPvEgp08cBxE6de9Dj/6D2L5lI/974XHiLl8mODiEF0a+Q5XqN3ub9c3pWK0QrSoVRIAZW47xy4ajNCmTl951i1Isb1YG/7CJncdjPRpDcgkJCdzWsC5FihThh1+mebXuJFXKlyZHjhwEBQUTEhJC1JIVXo/hi08/ZOL4cagq9/fpx0ODH/N6DGkJCoDMasYK8JLgkGCGDn+dmYvX8P2MRUwaP4bdO7bRsMkdTI9axbSFKylRugxffPS252IIDuHpl0by68JVTPx9AVO+Hcuendt57/WXeejxofwwewmDnnqB918f5rEYAErkzUqrSgUZ/MMmBkzewK0l81AkVxb2n7rI8Jk72HjoH4/Wn5rPPv6QyMjyPqnb0bRZ8/lrxRqfJNVtWzczcfw45kQtJWrZGubNnsnePbu9HkdqkrpbOVt8LdXEKiIf2dMRpLh4M8h/gwIFC1Opag0AsmfPQamykRw7ephGTZsTEmJ9cahesy5HD6c5aE6G5C9YiApVqgOQLXsOSpWJ5PjRw4gI52NiADgf8w/5CxbyWAwAxfJmZfvR81yOTyRRYeOhf2hcOi8Hzlwk+uwlj9admkPR0cyZPZNeffv7pH5/sXPHdmrWrkNYWBghISE0aNSEGVN/83VY1wkS54uvpdUUsDqNbUYGRB/8m22bNlCtZp3ryn+e8i0t297rlRgOHfyb7Vs2UqVGbZ4d/iYP9+zAuyNfIjExkW9/nefRuvefukj/W4uRM0sIl+MTqVc8DzuOn/donc4MfeYJXh05ivPnY3wah4jQoU1LRIS+/QfQp/8Ar9ZfoUIlXn9lGKdPnSJL1qzMnzOLajVreTUGZwL64pWqfuO4LiJhqurboZwySESqA0VUdaaT/ZoCT6tqa3fHEBt7niH9u/PCq6PJniPn1fLP3h9NcHAIbe/t5u4qb3Ah9jxPPdiTZ4aPInuOnHz81ms8M+wNmrdqx5xpvzDimUcYM3mqx+o/cOYiU9Ye5s12Fbh0JZHdJ2NJ9OFgTLNnTid/gQLUqFmLPxdH+S4QYPb8PygSHs6J48dp36YFZSMjadioidfqL1e+Ao8+8TSd27ckLCwblatWIzg42Gv1uyIA8qrzNlYRqS8iW4Ht9no1EfnU45F5RnWs4b984sqVKwzp3502Hbty1z3trpb/8v0EoubP4u1Pxnn8r/GVK1d48sH7adWhC81btgVg2s+TaWY/vqt1BzZvWOPRGABmbT3Ow99v4olftnD+cjzRZy96vM7ULF+2lFnTp1ElshT9enVncdQiBvTt6ZNYioRbM37kL1CA1m3asXb1Kq/HcH/vfiz4cyXT5iwiV+48lC5T1usxpEaAYBGni6+5cvHqfeBu4BSAqm4AvPcnNBkRKSEi20VkvIjsFJHvRKS5iCyxp62tKyLZ7OltV4rIOhFpJyKhwKtAVxFZLyJd7X2X2fssFZFIT8Wtqrz45MOUKhtJ34eGXC1fvHAuX37yPp+N/4GsYWGeqv5qDCOeGUypMpH0GvDI1fL8BQuxevlfAKxc8gfFSpT2aBwAubNaX5YKZA+lUelbWLDjpMfrTM2I115n254DbNqxl3HfTqJJ09sZ+/UEr8cRGxtLjN3WHRsby6IF86hQsZLX4zhxwho4P/rgAWZM/Y17O6c2wp4PuDCRoD80FbjU3UpVDyYLNsEz4bisDNAZ6AesAroDjYC2wAvAVmChqvYTkdzASmA+MAyoraqPAIhITqCxqsaLSHPgdSDNRk4RGYg1BS5Fwoumtet11qxcxu8/TaZchUq0a34rAE8+P4L/vfQMcXGX6dutDQDVatbl1dGeuTa4btVypv8yhbLlK9GlRUMAHn12GMNGfcToEc+RkBBPaObMDBv1gUfqdzSiVSQ5s4QQn6h8GLWX2LgEGpbKy6O3lSBX1ky83qY8u09cYOjUbR6PxV+cOH6MHt06AZAQH0+nLt1oflfyGZo9r2+PLpw5fZpMmUJ4890PyZXbM93/0ssP8qZTriTWgyLSAFARyQQ8Bvj6075PVTcBiMgWrLnAVUQ2ASWw5qRpKyJP2/tnAYqlcJxcwDciUhZQwOnQ7/aEZGMAKler6XLLYO16Ddhx5MZ+mbc1894vTs269dlwIOWuTFNmLvZaHACP/7zlhrIle0+zZG/yGZpDe1oAACAASURBVIm9q3GTpjRu0tQndZcoWYolK9b6pG5H0+dG+TqEVAkQ7A+X/Z1wJbE+BHyANd3rYWAONzmatgdcdnic6LCeiPWaEoB7VXWH45NEJPncHa8Bi1S1g4iUAKI8EaxhGO7jD1/1nXGaWFX1JNDDC7G40xzgURF51D6TraGq64AYIIfDfrm4NttiHy/HaBjGTcrgZIFe40qvgFIiMk1ETojIcRH5XURKeSO4DHgN62v9Rrup4DW7fBFQMeniFTAaeENE1mFu7zWMgBAk4nRxRkSeEJEtIrJZRCaLSBYRKSkiK0Rkt4h8b1/wThdXkskk4BOgg73eDZgM+GRKTFXdD1R2WO+TyrYHU3juaaBOsuJyDo9fsveLwjQLGIZfyuhYASISDgwBKqrqRRH5ASuvtQLeU9UpIvI50B/4LF0xurBPmKpOUNV4e5mIdTHIMAzDqwS33dIaAmQVkRAgDDgC3AH8ZG//Bmif3jhTPWMVkbz2w1kiMhSYgnXlvCvWPNuGYRje5Xo/1Xwi4nhb/hi7Rw+qekhE3gYOABeBucAa4Kyqxtv7R2NdsE+XtJoC1mAl0qRX4fjVWoHn01upYRhGernYEnBSVVMc+1JE8gDtgJLAWeBHwK39HtMaK6CkOysyDMPIKDf1Y22O1Rf+BICI/AI0BHKLSIh91hrBtR5DN82lK+EiUhmoiEPbqqp+m95KDcMw0ssN/VgPALeKSBhWU0AzrNH8FgGdsJo9ewO/p7cCp4lVRIYDTbES60ygJfAXYBKrYRhel9G0qqorROQnYC0QD6zDuptyBjBFRP5nl32V3jpcOWPtBFQD1qlqXxEpCExMb4WGYRjpJeKeW1pVdTgwPFnxXqBuhg+Oa4n1oqomiki8PWjJccD10UcMwzDc6F9xSyuw2h4haixWT4HzwDKPRmUYhpGKAMirLo0VMMh++LmIzAZyqupGz4ZlGIZxI8G1W1Z9La0bBGqmtU1VfT++mY9dSUjkyDnfTH6XpF7pvM538oJfB/jkDufrtPjwL1+HAMC0QQ18HQJx8Ym+DgH1xHQ79iyt/i6tM9Z30timWLd/GYZheJUr9+H7Wlo3CNzuzUAMwzCcEf49F68MwzD8RkgAnLKaxGoYRsCwBro2Z6yGYRhuFQDXrlyaQUBE5H4RGWavFxMRt9ydYBiGcbOSpmdJa/E1V1orPgXqA0mTi8dgzShgGIbhVQKEiDhdfM2VpoB6qlrTnhcKVT2TkblgDMMwMsIP8qZTriTWKyISjNV3FRHJjzXNtGEYhleJi5MF+porTQEfAr8CBURkJNaQga97NKp/ocuXL/FQ5+b0b9eEPq0b8PWHowBYu3wxAzreTp82DXnjuUHEx8c7OZJ7zZ0zm6qVIqlUvgxvjR7l1bodJSQk0OjWWnTp2MZrdXapVYSJfWsxsV8tutS6NgtHp5pFmNy/NhP71WLQbd4d7/2zj9+nYZ1qNKpbnQF97+fSJe/c2ffkIwOpWjaCO+rXuFo2euQImjesxZ2N63Bfx1YcPXLYK7E4869oY1XV74BngTewJtxqr6o/ejqwf5vQ0My8O/43vvp9MV/++gcr/1rA5rUreWPoYIa9M5bx05ZQMLwoc36b4rWYEhISeHzIYH6fNot1G7fy45TJbNu61Wv1O/rs4w+JjCzvtfpK5QujbdXC9J+wjt5fr6Fh6byE585CzWK5aFzmFnqNX8P949YweVW012I6cvgQYz//hPmLl/PXyvUkJiTw60/fe6XuLvf15Lufpl1X9vCjTzJ/yRrm/bmK5ne34r3RI70SS1oECAkSp4uvudIroBhwAZgGTAVi7TLjJogIYdmyAxAff4X4+HiCgoPIlCmUoiXLAFC7QVMWz52W1mHcatXKlZQuXYaSpUoRGhpK567dmD4t3YOmp9uh6GjmzJ5Jr779vVZn8VvC2HIkhsvxiSQorDt4jqbl8tGhehEmrDjIlQTrRvczF654LSaA+Ph4Ll28SHx8PBcuXKBQ4SJeqffWho3JnSfPdWU5cua8+vhC7AW/6T8aCGesrrSxzuDapIJZsCbg2gFU8mBc/0oJCQkMvPcODh3YR4fu/ahQtRYJCfFs37SO8lVq8MecqRw/ku5pdm7a4cOHiIi4NrRueHgEK1eu8Fr9SYY+8wSvjhzF+fMxXqtz74lYHmxcgpxZQrgcn0iDUnnZdjSGonmyUi0iFw82LkFcQiIfL9rLtqPnvRJT4SLhDB7yBNUrliJLlqw0bdac25vd6ZW6UzPqtWH8NOU7cubMyY/T5vo0FsAahMUPEqczrjQFVFHVqvb/ZbFG2DbjsaZDcHAwX/32Bz9GbWLbxnXs27WdYe98ySejXuKhzs3Jmi07QcHBvg7Tq2bPnE7+AgWoUbOWV+v9+/RFJq6I5v0uVXivcxV2Hj9PoiohQULOLCEMmLiejxft47W2Fb0W09kzZ5g1YxprNu1i864DXIi9wA9TvvNa/SkZ+vKrrN6yhw6d7+PrsZ/5NBawJxMUcbr42k3fdWsPF+j7MeK8QEQ8cmdajpy5qFGvESv/XEClGnX46LsZfP7jfKrVbkDREqU9UWWKihQJJzr64NX1Q4eiCQ9P91Tq6bJ82VJmTZ9GlchS9OvVncVRixjQt6dX6p6+6Sj9vl3HoMkbiLkUz8HTFzkec5k/dp0EYNvRGFSV3FkzeSWeP6IWULx4CfLlz0+mTJlo3bY9q1b4xzlMx87dmDn1V1+HAVhnrM4WZ0Qkt4j8JCLbRWSbiNQXkbwiMk9Edtn/53F+pFRidCGAJx2Wp0VkEuCzy4MiUsL+YXxn/0B+EpEwERkmIqtEZLOIjBG7QUhEokTkAxFZb2+ra5dnE5FxIrJSRNaJSDu7vI+ITBWRhcACd8V99vRJYv45B8DlSxdZvTSKYqXKcubUCQDi4i4z+csPaNutj7uqdKp2nTrs3r2L/fv2ERcXx4/fT+Ge1m29Vj/AiNdeZ9ueA2zasZdx306iSdPbGfv1BK/UnSfMSpgFc2Smabl8zN12nMW7T1GzWG4AiubJSkhwEGcveqedNSKiKKtXreTChQuoKoujFlLOixf0ktu7Z9fVx3NmTaN0uUifxeJIRJwuLvgAmK2q5bHm9NsGDAUW2N/MF9jr6eLKGVkOh8fxWG2uP6e3QjeJBPqr6hIRGQcMAj5W1VcBRGQC0BrrghtAmKpWF5EmwDigMvAisFBV+9lTz6wUkfn2/jWBqqp6OnnFIjIQGAhQsEiEywGfOnGMN4YOJjEhgURN5PYW7Wlw+918Nno4y6LmoImJtL2vHzVvbZKOH0f6hISE8N4HH9PmnrtJSEigd59+VKz032k6H9muIrmyhhCfqLw9bzfnLycwfeNRXmxZjol9a3ElMZH/zdzhtXhq1alHm/YduaNRXUJCQqhSrRq9+g7wSt2D+vdk2ZLFnD51klqVSvH00JdZOG82e3btJCgoiPCixRj17sdeiSUtQsbbWEUkF9AE6AOgqnFAnH1y1dTe7RsgCnguXXVoGsN82zcGvKmqT6fn4J4gIiWAxapazF6/AxgCTMDqFhYG5AU+UtVRIhIFvKqqC+39DwBVgflYF+OSOo7mBe7Gaua4TVX7OoslsnJ1HfPzQre9tvTwlxkE/GHEejODwDWX/eD9aHl7fTasW+PWBs+i5avoU2OnOt3viSal/gZOOhSNUdUxACJSHWu6661YZ6trgMeAQ6qa295HgDNJ6zcrralZQlQ1XkQapufAHpb8r4FijWlQW1UPisgIrKSZ1v4C3Kuq152SiEg9INa94RqG4Q43ccZ6UlVrp7ItBOtb6aOqukJEPiDZ135VVRFJ9+QyabWxrrT/X2+3OfYUkY5JS3ordJNiIlLfftwd624wgJMikh3olGz/rgAi0gg4p6rngDnAow5tsTUwDMPvuaEfazQQrapJfQt/wkq0x0SksFWHFAaOpzdGV9pYswCnsOa4SjrTU+CX9FbqBjuAwXb76lbgMyAPsBk4CqxKtv8lexCZTEA/u+w14H1go4gEAfuw2mUNw/BTQsa7U6nqURE5KCKR9jfWZlh5ZCvQGxhl/5/uu2XSSqwFRORJrGSVlFCvxpbeCt0kXlXvT1b2kr2kZKKqPu5YoKoXgQeT76iq44HxbojRMAx3c98NAo8C39kj9e0F+mJ9g/9BRPoDfwNd0nvwtBJrMJCd6xNqEl8nVsMw/qPcMbqVqq4HUmqDbZbhg5N2Yj2S1H3Jn6jqfqzuUq7u39RjwRiG4VXWLK2+jsK5tBJrAIRvGMZ/TXAADBaQVmJ1yymxYRiGuwjpuA/fB1JNrCnddWQYhuFTZvprwzAM9/P/tGoSq2EYASRp2EB/ZxKrYRgBJQDyqkmshmEEEpeHBfQpk1gNwwgYpinAMAzDA/w/rZrEmiFZQ4OpUCSH8x09aM8x70x050xE3qy+DoGZj/jHCJf9Jq/3dQgMv7Ocr0MgPsEDd76b7laGYRjuFfA3CBiGYfgjdwzC4mkmsRqGEVACIK+axGoYRuCwmgL8P7OaxGoYRgAR0xRgGIbhbgGQV01iNQwjcJimAMMwDHdzbRZWnwuELmGGYRhXBYk4XVwhIsEisk5EptvrJUVkhYjsFpHv7YkG0xdjep9o3JzHBg2gYqlwmtSrfrXszOnTdG7XklurV6Rzu5acPXPGozEcORxN386taHt7bdrdUYcJX356ddt34z6nzW01aXdHHd75X2qT3XpGlfKlaVCnOo3q1aJpw3perTvJubNn6d2jC3VrVKJezcqsXLHMa3W3rlSA9ztW5P2OFXmiaUkyBV9LDP1vLcp3vaqn8eyMO3o4mge63kPHZnXo2Lwu342zPhc7tm6iV/tmdLrrVob068L5mH88GocrBGuWVmeLix4Dtjmsvwm8p6plgDNA//TGaRKrl3Tr0Yspv0y/ruyj90bT+LbbWb5+K41vu52P3hvt0RhCgkN4ZtjrTF20mklTFzLlmzHs2bmdlUsWs2juDH6eu4zfF66iz0OPeTSOlEybNZ+/VqwhaskKr9cNMPSZJ2h2592sXLeFP5evJTKyglfqzRuWiXsqFeDZ37fx+C9bCRJoVCovAKXzhZEtc7DHYwgODuGpl0byy4JVTPhtAd9/O5Y9O7fzynOPMGToK/w0dzl33N2Gb774wOOxuEJc+Of0GCIRwD3Al/a6AHcAP9m7fAO0T2+MJrF6Sf2GjcmdJ891ZbNnTKNr954AdO3ek1nTp3o0hvwFC1GxinX2ky17DkqVjeTY0cN8P+FL+g9+ktDMmQG4JV9+j8bhb86dO8fSJX/Ss3c/AEJDQ8mVO7fX6g8WITQ4iCCBzCFBnL4QR5BArzoRTFgZ7fH68xcsRAXHz0WZSI4fO8yBfXuoVc8af+HWxrezYJZnP5+uEnG+APlEZLXDMjDZYd4HngUS7fVbgLOqGm+vRwPh6Y3RJFYfOnHiOAULFQagQMFCnDhx3Gt1Hzr4N9s2b6Rqjdrs37ubNSuWcl/r2+lzbws2rV/jtTjAGlSjQ5uW3NagLuO/GuvVugEO7N9Hvnz5GPxgf5rUr82QQQOJjY31St2nL1zh983H+KJbFb66ryoX4hLYcCiGlhULsOrAWc5cjHd+EDc6dPBvtm/ZSJXqtSlVtjyL5s4AYN6M3zh65JBXY0lJ0rCBzhbgpKrWdljGXD2GSGvguKp67IMesIlVRKJEpHY6ntdeRCo6rL8qIs3dG93NE/HeAL4XYs/zxMD7eW7EKLLnyElCQjz/nD3DpGkLeeql//H0w71R9cDIRKmYPf8PFi9bxU+/TWfsmM9Y8tdir9UNEJ8Qz4b16+g34EEWL1tNWFg23n/nTa/UnS00mLrFcvHwD5t5YPJGMmcKpmmZvDQokYeZW733hxasz8XTD/XkmWHW5+KVtz7lhwljue+eJsTGxpApUyavxpMyVxoCnP4eNQTaish+YApWE8AHQG4RSeopFQGk+y9JwCbWDGgPXE2sqjpMVef7IpD8+Qtw7OgRAI4dPUI+L3wFv3LlCo8PvJ97OnThzlbtAChYKJzmLdsiIlSpURsJCuLM6ZMejyVJkXDrG1f+AgVo3aYda1ev8lrdAEWKRFAkPILadawLZ207dGTD+nVeqbtqkRwci4njn0vxJCis2H+GrjWLUChnZj7tXJnPu1Qmc0gQn3Su5NE4rly5wlMP3U+r9l1o1rItACXLlOPzib8zecZiWrbtRETxkh6NwSUuNAM4Oz9R1edVNUJVSwDdgIWq2gNYBHSyd+sN/J7eMP0+sYpICRHZLiLficg2EflJRMKS7fOZ3Y6yRURecSgfJSJbRWSjiLwtIg2AtsBbIrJeREqLyHgR6WTvX0dElorIBhFZKSIeHWz17lZt+H7SBAC+nzSBFve08WR1qCrDnh5MqTKR9B746NXyO1q0ZuVS6yxx/95dXImLI0/efB6NJUlsbCwxMTFXHy9aMI8KFT2bRJIrWKgQ4RER7Nq5A4DFUQuJLO+di1cnY+MoVyAboXZPgCpFcjJt8zH6T97IQz9s5qEfNnM5PpHBP27xWAyqyivPDqZkmUh6DnjkavnpkycASExMZOxHb9G5R7ovkrvNTTQFpMdzwJMishurzfWr9B4oUG4QiAT6q+oSERkHDEq2/UVVPS0iwcACEamKdRrfASivqioiuVX1rIhMBaar6k9wbdBcu8/a90BXVV0lIjmBi8kDsRvBBwJEFC3m8gt4sO/9LP1rMadPnaR6+ZI888IwHn3iGQb06c6kb8cTUawYY8dPurmfyk1at2oZ036eTNnylbj3rgYAPPbccDp27clLTw2ifbO6ZMoUyuvvf+G1ZokTx4/Ro5t1kpAQH0+nLt1oflcLr9TtaPTbHzCwXy/i4uIoUbIkn3ye7t+pm7LrxAWW7TvD2+0rkqjK3lMXmLvde98WANavXs70X6ZQtnwlurS0LlY9+swwDuzfw/ffWm3ezVq0pV2X+70aV2rc+clU1Sggyn68F6jrjuOKN9vS0kNESgCLVbWYvX4HMATIDTytqqtF5CGsZBcCFAYexeo2scZepmMl0zgRGc/1iXW8vX0H8LmqujwMffWatXTuH8vd8CrT78Q/l31afxJ/mEHAXz7KZgYBS/fWt7Fl41q3/oWuUKWGfv3bIqf71S+TZ42q3vQ1GHfx+6YAW/JfmavrIlISeBpopqpVgRlAFrvbRF2sBNsamO2lWA3D8CB39GP1tEBJrMVEpL79uDvwl8O2nEAscE5ECgItAUQkO5BLVWcCTwDV7P1jgJTaTncAhUWkjv38HA5XCA3D8BNuvPPKczH6OgAX7QAGi8g2IA/wWdIGVd0ArAO2A5OAJfamHMB0EdmIlYiftMunAM/Y9wiXdjhOHNAV+EhENgDzgCwefVWGYdw8cWHxsUA5I4tX1eQt502THqhqn1Sed0NDtKouwaG7FdDHYdsq4Nb0BmkYhmdZedMPMqcTgZJYDcMwwE++6jvj94lVVfcDlX0dh2EYfsIkVsMwDHfyj6v+zpjEahhGQAmEGQRMYjUMI2AIJrEahmG4nWkKMAzDcDNzxmoYhuFmAZBXTWI1DCOACF4beS0jTGI1DCNgmItX/wHxCcrJmDifxlD0ljDnO3nB+UvenZspJVky+cfQF083KeXrEPh0xQFfh8DxWM8MaRkAedUkVsMwAotpCjAMw3CzAMirATNsoGEYBpDxUQNFpKiILLLnw9siIo/Z5XlFZJ6I7LL/z5PeGE1iNQwjsGR8PNZ44ClVrYg1TOhgEakIDAUWqGpZYIG9ni4msRqGETBEIEjE6ZIWVT2iqmvtxzHANiAcaAd8Y+/2DdA+vXGaNlbDMAKKi02s+URktcP6GFUdc8OxrMlKawArgIKqesTedBQomN4YTWI1DCOwuJZZTzqbpdWeF+9n4HFV/cext4Gqqoike95f0xRgGEYAcWWOVueZV0QyYSXV71T1F7v4mIgUtrcXBo6nN0qTWA3DCBhCxmdpFevU9Ctgm6q+67BpKtDbftwb+D29cZqmAC85ciia5x8bwKmTxxEROvfoS88HBvPJOyP5adJ48uTNB8DjQ0fQpNndXosrISGB2xrWpUiRIvzwyzSv1PnkIwOZP2cm+fLlZ+GydQCMHjmCuTOnIUFB5Mufn/c++ZJChYt4JR6AXTt38EDv7lfX9+/fx/MvDeehwY95tN7Lly8xuHtrrsRdJj4hntvvbssDjz3Pw/e14kLseQDOnD5JxSo1GfXZRI/EUDB7KAPqR1xdz5ctlGlbjpM1UzCNSuXh/GXrrrrfNh1n89HzHonhpmS8H2tDoCewSUTW22UvAKOAH0SkP/A30CW9FZjE6iUhISE8O/wNKlapTuz5GDq3aEz9JncA0GvAI/R9yLO/wKn57OMPiYwsT0zMP16rs8t9Pek74GEee6jf1bKHH32SZ18cAcBXX3zMe6NH8uZ7n3gtprLlIvlj2RrA+mNTuWxx7mmT7ovCLgsNzcyH3/5GWLbsxF+5wsP3teTW25rz2eSZV/d54ZFeNG7WymMxHDsfx//m7QWsnPVmm3KsOxRDgxK5WbDzFPN2nvJY3emR0fFYVfUvUk/PzTJ0cJtpCvCS/AULUbFKdQCyZc9BqbKRHD96xMmzPOtQdDRzZs+kV9/+Xq331oaNyZ3n+r7XOXLmvPr4QuwFn962uDhqISVKlaJoseIer0tECMuWHYD4+CvEx8df99pjz//D2uV/0uROzyVWR+ULZuPE+SucvnDFK/WlR0abArwSo68D+C86dPBvtm3eQNUa1kXLSV9/QYfm9XjpyYc5d/aM1+IY+swTvDpyFEFB/vExGPXaMGpXKs2vP07mmReG+yyOX376no6dunqtvoSEBHq3bULr+pHUadiUStWuXcxePG8mteo3IVv2nGkcwX3qFM3FqgPnrq43LZOXl+8sTa/aRQjzh0FuxOrL6mzxNT/4SaWPiOQWkUFuOlaUiKTZNcNdYmPP8/iAHgx95U2y58hJ114PMHvpJn6eu4z8BQry1qsveCMMZs+cTv4CBahRs5ZX6nPF0JdfZfWWPXTofB9fj/3MJzHExcUxe8Z02nXo5LU6g4OD+WbqYn5dvJmtG9eyd+fWq9vmT/+Z5q3v9U4cIlQrkoM10VZi/WPPaV6auYv/zdvDuUvxdKpWyCtxOJfxW688LWATK5AbuCGxiojfthtfuXKFxwf04J4OXbmzVTsA8uUvSHBwMEFBQXTq0ZdN61c7OYp7LF+2lFnTp1ElshT9enVncdQiBvTt6ZW6nenYuRszp/7qk7rnz51N1eo1KFAw3X3D0y1HzlzUrNeI5X8uAODs6VNs3bSWBk3v8kr9lQtn58CZS8RcTgAg5nICCijw194zlMib1StxpCVpPFZzxuo5o4DSIrJeRFaJyJ8iMhXYCiAiv4nIGnuQhYF2WbCIjBeRzSKySUSecDheZxFZKSI7RaSxu4NVVYY9NYhSZSLp8+CjV8tPHDt69fH8WdMoG1nR3VWnaMRrr7NtzwE27djLuG8n0aTp7Yz9eoJX6k7J3j27rj6eM2sapctF+iSOX378no6dvdcMcOb0SWL+sc4QL1+6yKolURQvVQ6ARXOm0qDp3WTOnMUrsdQpmotVB681A+TMcu0cpXp4Tg6f88z4qjcrENpY/fbszgVDgcqqWl1EmgIz7PV99vZ+qnpaRLICq0TkZ6AEEK6qlcFqTnA4Xoiq1hWRVsBwoHlKldpJeiBA4fCiLge7dtUypv48mXIVKtHxzvqA1bVq5m8/sn3rRkSEIhHFGfHmhy4fM1AN6t+TZUsWc/rUSWpVKsXTQ19m4bzZ7Nm1k6CgIMKLFmPUux97Pa7Y2FiiFs3n3Q8/9Vqdp44f43/PDSIxMYHExETuaNmehrdb3e0WzPiF+wd6p7dIaLBQoWA2Jq45fLXs3qoFKZo7C6pw6kIcE9f49mJrkkCYpVVU033Xlk/Z9/hOV9XKdmIdrqq3O2wfAXSwV0sAdwM7gNXATKxEPFdVE0UkCnhRVZeISEFgiaqWcRZD5Wo19YdZf7rrJaVLhB98PQMzg4Cj7YdjfB0C36w/7HwnD/tlaBdO7Nni1ixYrUYtnfPHcqf7Fc4VusbZLa2e5B+fRPeITXpgJ9rmQH1VrQasA7Ko6hmgGhAFPAR86fD8pO85CQT2mbxh/GuJC80ApikgY2KAHKlsywWcUdULIlIea8xFRCQfEKeqP4vIDsAzt7IYhuExgdAUELCJVVVPicgSEdkMXASOOWyeDTwkItuwvv4nfXcIB74WkaQz9ee9FrBhGO7h/3k1cBMrgKp2T6X8MtAylafVTGH/pg6PT2K1yRqG4YcCIK8GdmI1DOO/xvkMAf7AJFbDMAJG0g0C/u7f1CvAMAzDL5gzVsMwAkognLGaxGoYRuCwZ2n1dyaxGoYRMPxj7CrnTGI1DCOwBEBmNYnVMIyAEghNAaZXgGEYAcUdw1yLSAsR2SEiu0VkqLtjNInVMIzAksHMKiLBwCdYd2dWBO4TEbcOhGwSq2EYAUVc+OdEXWC3qu5V1ThgCtDOrTEG6nis/kBETmDNP54R+YCTbgjHxOAe/hCHP8QAGY+juKrmd1cwACIyGysuZ7IAlxzWx6jqGPsYnYAWqvqAvd4TqKeqj7grTnPxKgPc8aERkdW+HJDXxOB/cfhDDP4UhyNVbeHrGFxhmgIMw/ivOQQ4zqsUYZe5jUmshmH816wCyopISREJBboBU91ZgWkK8L0xvg4AE4Mjf4jDH2IA/4nDrVQ1XkQeAeYAwcA4Vd3izjrMxSvDMAw3M00BhmEYbmYSq2EYhpuZxBqgHCZENLxEJABuUjf8gvnlDCAiUlVEfgVQ1cR/e3IVkdK+jiGZMF8HYASGf/Uv5r/QfuCKiEwCzyTXpLMyESnpzuOmI45cwEgRGenLOJKISEvgexHJ6qszV4f3pogv6jVcZxJrAEj6YKvqxW3z9gAAEatJREFUP8BnWH3wvrbL3JpcVVVFpBUwV0Qi3XXcdIjFGiijmIgM82EciEhl4AlgmKpeVB90pRERsd+blsAsESnhjYSXVK/9uIiI5HPc5un6A5VJrAHA4YP9LDAUmAmUEpGf7e1uS64iUg94F7hPVXeIyC0ikskdx3ax/qQ/IvHACuBToJKvkquIZAc6AFWxmwK8mVAcfh4qIrWBt4Ahqrrf0wleRGoA1f/f3plHW1Fdefj7KaMKiAg02kGNgMSQiMYozkCrqIigKKiIwREnMOJApzWKJu2QGDHGtKiAQzC2orZjHBJbBQda1AiKxjibOA/gPET55Y997uL2W6i896rue8L51qr1qureW+dU1atd++zppPUJwM3AdElHV/UpC9elkAVrM6b6n1ZSG2BrYJztU4hqPJ9Xa64FNbsqcCnQTtKPgbuAKZJ6FXT8L6WOdrQ6gO0HgMlAn1oLV0mtbH8ATAEuB/aX9P1aaaxpyL9DKnMH0B64zvY9klrVwMY+HDhD0nCgHzASOB04RNLxsOSln/n/ZMHaTKkjZEYA2xMPVsWh8x5wLTBI0oWNaSf93UbSFsDfgc2BicBCYH8iO6V0s0DV+Y4nBNkUSaNszyG06N6Sziy7H6kPQ4GpkmYCPYHfEZXMDk6aXC3YhLgfq0lqD3xBCPcetj9LI5XtU18LoyKwbZ8E3ANMAF6z/XS6F/sC+0k6pch2lyts56UZL0B/YFZa3xd4BuiXtvcDTibKszWmjWHAn4FBabsjsEpa75k+61ej8z0MmAWsDVwBvAIcnT7bGpgOrFlyH7YG5gDdgFuAm9P+3sDPCQ22XYntr1y13hGYQbzgWgFHE6mY2wADgMeAXQpsW1Xr66e/xxIjl40qfSOE/hygU/Vv8pKuXVN3IC9fcXNgR0JjOLVq3yGEFnMh8BLQq5Ft/AtwH/CvaXtDYNe0vjPwEDCsxHNsW7W+JjHc7JQEyPVJgDxH2BUB2tTguh9I2FWHAvcD66X9rYhKSD3LvB7ApkQ90W2A76V7fkm6NusCRwB3AzdU7lUJ/TgGmAl0TtuTgJuAjauEa+uy78U3dcm1ApoR1cP/tL064RlvCxxl+5W0vy8xLHzPdqMKbUtah3hApxMaWUdCoBxOaEZr2X6kbt+KQNKqwL8BbwDrAS2Jh7c1IUgOtf03SdcQpohtbS8ssg9f0q/hhDBbBRhj+zlJ+xDmmLEOx1pZba9JCM4NgG0JwTlP0hhCQ73F9tXpu61sf1b0vUmmpx8TxaDfk9TO9vuSjiAqQR1le35R7S2P5OpWzYQ6NtUBwAfA48DBhNA7QdIvbL9i+9HGtiOpO/C67RclTQL2Aq6w/Ydks9va9mXAa1Cqk2Ix4flfHdjE9iJJ3YBFROTDLsDbwMAyhWq65msBnxMvmrGEtvq5pO2AE4GJJQtV2X5L0n2Exn4NaYYK25dKWgyMTM6saxzTipRxb7oTw/zeKbyrv6R/2N4xRUm8U3B7yx1ZY21mSDqWsHk+TQid3wJ/IUq4fQqcaPv1Bh57JYfDYzfg+HTcBcAltt9N39mB8MIfa/v2xp7Pl/Sj+iXyA2LIfz8wxfZdaf/JhJ11C2C07Xll9CW1tQVwFWE73Z8YAl8EnEYM/zsB59u+uQzNPfWh8sLblIjM+Jywqb8M3OBU1k7S/sC8oq5HnXvRxvYnknoA5xEjpWnAA4Qz83TbLxTR7nJPU9si8rJkAXYC/pDWJwOPEA97X2JYejHQtQHHbV21vhXwMGFb/S3wKPCfQA9CgMwFhpR4jtXOkW7E8H81wn54MbBn+my9tKxW8jX/IXA+MCJttyG0tUmV/pKcZZTspCFeqP9HmDwg7KuXACcA44j45dVLuheHA2cDxxGmGLFE8dqDeAF3K/P8l6clh1s1IUsJrn4VOFLSwUAfYE9ieHoW0Mf2Ia6ntiqpCzBRUr+0qxNwKOHV3ZwQqpsARxGmoUG2byor8NuVJ1o6jhCkswhzx1zgQWBnSTMI59z7jjjSwqk6vy2IF1rvZEv8hIjf7CepQ+rv29V9L6k/3ybMDXvZniVpXcKOPoF4qe5AFGReVFSbVffiSEI7npr6cC7wbaBNsjefTrx4Xi2q7eWdbGNtIuoMwXoRYUXP2P5QMcf5GQ6nyXxCo3uhgU0tBtYBdpX0se0bk1CZAhxg+zFJgwinVQfbf4XShcgwYHvbO0m6HOhv+1xJ7xIRACOBCbbLnKl0TeBN2+dJWgSMBmZLepCwMXYh4ndLvRZVtCbu1Q6SehLOq0HAvrYnSVrF9kdFmCLqDv+BLYmX+F7EKKkLEVZ2CjAf2NH2S41pc0UjC9YmouofexwwgnASvSXpV8CTwNWSziGGh7vZfqO+bUha2eEMOZ/QSPeQ1NL2Q4oiJ6dK+gURn3h4RagWTcW2W7XrQ2CapJ8CXYEhaX8n23cCd5bRj6r+7AqMk/QYcJ/tyxVzH51FvMBaAqfZLs1JU2VT7Qwssv2kItFjNPBrQnMcTmiO2P4o/S1SqE4kTEEHEim7w20PkNSJuA6PApOdnGSZZScL1iZEkZd/CBHG051IGzwdOIjQXjYghmDPNeDYsv2FIktob+ApIth/iKQPiOH3VOBUwinxUAGntFQqQjVpqh8Rdt6NCDveYMccROOILLIRQGmFTiT1J8wfwwlB2k/St5Lm+iEwHvil7evLaD/1oSJUhxJmmXckzbQ9nYgAqfRzUupPYVQJ1Z2J5JMrbX8q6VOghSKNdmPCnntlFqoNIwvWGlL1QFW0hjbAE0kbfUPSa4Qzpa/taZJauIHhPamdISxxRrxEOK16EYL2Wtt7Jzviu2V4u+toR3sTDrmLiSFuVyKcaLdkTxxDFH75qMg+LIXexPlvQJhIpgHDFGmcvybCvo6V9Cowp8hrUtHc070ZQGTN7UrYNE9VhJndSNyvyUR4158KarsLEey/QBET+++E6akyxH+VSDqYQTg2h+fhf8PJzqvaUnGYtEt//wysm5wH2P474bDomT7/osENSV2JEJmxtvsRGVztiNjY9YF9JLV3CrMqWah2B0zExp5MaOXvEhlGXYhsr5EueKbMOv3ZOmnDzxF1FgYTwuOC1Le+RPbZBYRweblgobo24UTsmnb1INJ3f0BkU00lXi4jgI+JNNUinYgdgMmSLgN+RGjt3RR1GbD9JlGP4UjCgflkQe2ukGSNtQYkj/yrjmD88YSWtoAQrEcDP5HUG5hHPGg/h0YLu8+IF2entH0REYjfh8ioutdR37Vw6gjV8cAoQqifI+ll2zckgfEb4GHbU8roR1V/tiQ05bnEy6oDEQnxiKQHCOfgZNt/A0jCtWgWA7sAbRWJHhdLWoXIw98nOSp3JCI1rqvqSyHC3fbTyRF6KKEJ/07SW8BYSdg+z5GAUXpm24pA1lhrw2DgZkk/IlI4f0Y4BvoT9QDGpe91B0Y1xKZal/SQXAsMlNTH9j+IoXdrQqg81dg2vqLtilAdRmilo4lant8jbJotkg3zeFIoU1lI2ozQzg6wvT9x7e8iaiAcQZgCznG5CQgtUqhSpa7rhOS0+phIQBinSAzoAJxbEaolMIXQSA+VNNL2rcT1GKVI2c0UhZtBMO3yuhD20o5pfSKRnjghba9GhLlcQdj1Cg8+JwqG/Ay4lRAuTwGbEXa8jUo+97UJu+60tN0m9eU3RExmixrdgx0ILfU/0nZLohbC6ekedEn7yw7+346wp44GZhOaakfiZTOT0KaH1uiaDCHCqAalazGbVGgmLwVd46buwPK8EAWjHyJlyxA2rNeBtdP2SsBtwMYl9qE9EQA/MT3Em6eHqt4ZXA1oew8iPneftN0iXYOzSWUJa3QfhhLlFiv92I5w5JVdfrCSubQlEUI3g5gB4G4iw+oYlmR1dav+TQ2uyU5EycE5wHdrdS9WlCXXCiiBFD/6RVq/nhAoo20vVEyON4wIo2lFCJmdXN7wr7pfA4AzCIdWaUPfOm0OTm2eYftKSS0ILf7NWrRf1Y8hxOjgDsLeOcP2jTVodzMirOsntuco8vAHE86yHqQ8fDfBg5giBVzre7EikJ1XJVAlVMcAbxJ21bsl9bd9oqS2wB+JB333WgjVxF8I73ujSg3WB9u3KKoyXSTpc9sziWtSUxwe9v2IwipXeEkGGiULtQ5E+b+BhHb4IvA8EdJ0EjFyaBLtxg1IOsksG1ljLZA63vB+wO+JXPQ3iGIa3yXSAxemEKtbXYCj6puAomrWs019vsnzPp0onH1djdocCvwK+GnS2rcj4lQHusDc/0zzIQvWgqgjVDcm6omeRKSKfpb2309oMFvlB6rpaAoh31SmiEzTkMOtCqJKqI4BziQyi9YnQqoqzCAC41etcfcyVdj+Y601Z9s3EXOU9QDmVkwRBSYAZJoR2cZaIJK2IiozHW97vqQLgLNSTYD2RC2AEbZfbsp+ZpqGJEw/AaZLerZWpohM7cmCtRFU5f6vRGj/mwKdgeGSnrJ9Vcr/70VkHo11pK1mVlBs3yHpAODZpu5LpjyyjbWB1LGpdnUqQC1pNBErej9wtUucIymTyTRPssbaQKqE6pHAUEnzgMdtXyapJSFcW0u6vBJ+lclkVgyy86qeVDsbkqNqX6KwxTpEDvgJjrqafwW+Q3ZUZTIrHFljrQd1hv+bAu8T+d+jCOfUeMJZtdj22anWaSkVpDKZTPMla6z1oEqoHk5MnTGPeDltD+xn+x4iN76/pDWcap1mMpkVi6yx1hNJuxFTBQ9x1FftRmirvRRzKS0GxrjE+ZIymUzzJgvW+rMW8N9JqLa0/aqkW4iaqt2BI13u7KKZTKaZk00B9edFYFtJGziKR0PUOb2dmNL50abrWiaTaQ7kONZ6Iqk9Ufm+BXAfUaT6aKLW5zNN2bdMJtM8yIK1ASS76lBgNyL3/wzb85u2V5lMprmQBWsjkNQKwHnu9UwmU0UWrJlMJlMw2XmVyWQyBZMFayaTyRRMFqyZTCZTMFmwZjKZTMFkwZrJZDIFkwVrpsFI+kLSo5IelzRT0iqNONalkvZM61MlbfgV3+0vacsGtPGCpDWXdX+d73xQz7YmSTquvn3MLB9kwZppDB/b7mu7D/AZcFj1h5IaVIvC9sG2n/iKr/QH6i1YM5lakQVrpihmAz2SNjlb0o3AE5JWlvRLSXMlzZc0FqK2raTzJT0l6U9Al8qBJN2d6t0iaSdJj0iaJ+lOSesSAvyYpC1vI6mzpGtTG3PTpI5I6iTpDkkLJE0FvnZGVEnXS3o4/ebQOp9NTvvvlNQ57Vtf0m3pN7Ml9S7iYma+2eTqVplGkzTTnYHb0q5NgD62n0/C6V3bP5TUGrhP0h3AxsAGwIbEVOFPANPrHLczcDGwbTrWGrbfkTQF+MD22el7vwcm275XUneiIM53gFOAe22fJmkwcNAynM6BqY22wFxJ19p+m5gJ4iHbx0g6OR37KOAi4DDbT6fZeP8LGNiAy5hZjsiCNdMY2kqqVPOaDUwjhugP2n4+7d8R+H7Ffgp0AHoC2wJXpvnAXpH0v0s5fj9gVuVYX1Hjdntgw6pZc9pLWi21sUf67S2SFi7DOY2XtHta/1bq69tEnd2r0v4ZwHWpjS2BmVVtt16GNjLLOVmwZhrDx7b7Vu9IAubD6l3AONu31/neLgX2YyWgn+1PltKXZUZSf0JIb2H7I0l3A22+5OtO7S6qew0ymWxjzZTN7cDhaeZaJPWStCowCxiZbLDdgAFL+e0covbteum3a6T97wPtqr53B1FonPS9iqCbRUz2iKSdgY5f09cOwMIkVHsTGnOFlYCK1r0vYWJ4D3he0l6pDUna6GvayKwAZMGaKZuphP30EUmPAxcSI6X/AZ5On10OPFD3h7bfJGbAvU4xvXhlKH4TsHvFeUVM4rhpco49wZLohFMJwbyAMAm89DV9vQ1oIelJ4ExCsFf4ENgsncNA4LS0fxRwUOrfAqKcZGYFJ1e3ymQymYLJGmsmk8kUTBasmUwmUzBZsGYymUzBZMGayWQyBZMFayaTyRRMFqyZTCZTMFmwZjKZTMH8EyGF2MMJy0PNAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}