Convolutional_Network.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "23iDagpAzbj0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "af90f48f-524f-4841-dd38-893ed5c68335"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EoCWHk2yZqhw"
      },
      "source": [
        "### Unzip"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "muxrM_LYXl8i"
      },
      "outputs": [],
      "source": [
        "!unzip /content/drive/MyDrive/blockchain/Brain_Tumor.zip >& /dev/null"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "iIL07WNj6lMd"
      },
      "outputs": [],
      "source": [
        "!unzip /content/drive/MyDrive/blockchain/Dataset.zip >& /dev/null"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rPS9epThrsu6"
      },
      "source": [
        "# Libraries"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install tensorflow_addons"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hLzpeGJkTWRB",
        "outputId": "9fa434a2-642c-4e9f-e0d5-b56efe90e28e"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting tensorflow_addons\n",
            "  Downloading tensorflow_addons-0.23.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (611 kB)\n",
            "\u001b[?25l     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/611.8 kB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K     \u001b[91m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[90m╺\u001b[0m\u001b[90m━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m276.5/611.8 kB\u001b[0m \u001b[31m8.1 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m611.8/611.8 kB\u001b[0m \u001b[31m11.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from tensorflow_addons) (23.2)\n",
            "Collecting typeguard<3.0.0,>=2.7 (from tensorflow_addons)\n",
            "  Downloading typeguard-2.13.3-py3-none-any.whl (17 kB)\n",
            "Installing collected packages: typeguard, tensorflow_addons\n",
            "Successfully installed tensorflow_addons-0.23.0 typeguard-2.13.3\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "O90KzDOmp8zT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0ca7b6e5-b29d-4c3b-916a-710b9ff27a7b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/tensorflow_addons/utils/tfa_eol_msg.py:23: UserWarning: \n",
            "\n",
            "TensorFlow Addons (TFA) has ended development and introduction of new features.\n",
            "TFA has entered a minimal maintenance and release mode until a planned end of life in May 2024.\n",
            "Please modify downstream libraries to take dependencies from other repositories in our TensorFlow community (e.g. Keras, Keras-CV, and Keras-NLP). \n",
            "\n",
            "For more information see: https://github.com/tensorflow/addons/issues/2807 \n",
            "\n",
            "  warnings.warn(\n"
          ]
        }
      ],
      "source": [
        "import random\n",
        "import numpy as np\n",
        "from  drive.MyDrive.blockchain.classHospital import Hospital\n",
        "import os\n",
        "import cv2\n",
        "import pickle\n",
        "import pandas as pd\n",
        "import gc\n",
        "from sklearn.model_selection import train_test_split\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.constraints import MaxNorm\n",
        "from tensorflow.keras.initializers import GlorotUniform, Zeros\n",
        "from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, Flatten, Dense, Dropout, Rescaling\n",
        "from sklearn.metrics import classification_report, f1_score, accuracy_score\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.models import load_model\n",
        "from tensorflow.keras.callbacks import ModelCheckpoint\n",
        "from sklearn.metrics import f1_score\n",
        "import tensorflow_addons as tfa"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9V4HwQ5Jq1cA"
      },
      "source": [
        "# Constants"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "wUQbZsrUq8F2"
      },
      "outputs": [],
      "source": [
        "BRAIN_TUMOR = \"Brain_Tumor\"\n",
        "ALZHEIMER = \"Alzheimer\"\n",
        "\n",
        "\n",
        "RANDOM_SEED = 42\n",
        "LABELS_ALZ = [\"NonDemented\", \"VeryMildDemented\", \"MildDemented\", \"ModerateDemented\"]\n",
        "LABELS_TUMOR = [\"glioma\", \"meningioma\", \"notumor\", \"pituitary\"]\n",
        "NUM_CLASSES = 4\n",
        "COMPILE_INFO = {\n",
        "    \"optimizer\": \"Adam\",\n",
        "    \"metrics\": [\"accuracy\"],\n",
        "}\n",
        "WIDTH = 176\n",
        "HEIGHT = 208\n",
        "DEPTH = 1\n",
        "BATCH_SIZE = 32\n",
        "DATASET_TRAIN_PATH_ALZ = \"/content/Dataset/train\"\n",
        "DATASET_TEST_PATH_ALZ = \"/content/Dataset/test\"\n",
        "DATASET_TRAIN_PATH_TUM = \"/content/Brain_Tumor/train\"\n",
        "DATASET_TEST_PATH_TUM = \"/content/Brain_Tumor/test\"\n",
        "\n",
        "train_path = DATASET_TRAIN_PATH_TUM\n",
        "test_path = DATASET_TEST_PATH_TUM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "XL6kDqNigg3j"
      },
      "outputs": [],
      "source": [
        "def set_riproducibility():\n",
        "  os.environ[\"PYTHONHASHSEED\"] = str(RANDOM_SEED)\n",
        "  random.seed(RANDOM_SEED)\n",
        "  np.random.seed(RANDOM_SEED)\n",
        "  tf.random.set_seed(RANDOM_SEED)\n",
        "  os.environ[\"TF_DETERMINISTIC_OPS\"] = \"1\"\n",
        "  os.environ[\"TF_CUDNN_DETERMINISTIC\"] = \"1\"\n",
        "  tf.keras.utils.set_random_seed(RANDOM_SEED)\n",
        "set_riproducibility()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k5V2Cn8trP7j"
      },
      "source": [
        "# Setup"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "Vkuqh33ashoM"
      },
      "outputs": [],
      "source": [
        "def preprocess_data(image, label):\n",
        "    label = tf.one_hot(label, depth=NUM_CLASSES)\n",
        "    return image, label\n",
        "\n",
        "def count_classes(dataset):\n",
        "    class_names = LABELS_ALZ if train_path == DATASET_TRAIN_PATH_ALZ else LABELS_TUMOR\n",
        "    class_counts = [0] * len(class_names)\n",
        "    for _, labels in dataset:\n",
        "        class_counts += np.sum(labels.numpy(), axis=0)\n",
        "    return class_counts\n",
        "\n",
        "def create_dataset(train_path, test_path, dataset_name):\n",
        "\n",
        "    labels = LABELS_ALZ if train_path == DATASET_TRAIN_PATH_ALZ else LABELS_TUMOR\n",
        "\n",
        "    test_dataset = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "      test_path,\n",
        "      seed=RANDOM_SEED,\n",
        "      image_size=(WIDTH,HEIGHT),\n",
        "      batch_size=BATCH_SIZE,\n",
        "      class_names = labels,\n",
        "      color_mode='grayscale',\n",
        "      shuffle = False\n",
        "      )\n",
        "\n",
        "    train_dataset = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "      train_path,\n",
        "      seed=RANDOM_SEED,\n",
        "      image_size=(WIDTH,HEIGHT),\n",
        "      batch_size= BATCH_SIZE, # it is needed for the split\n",
        "      class_names = labels,\n",
        "      color_mode='grayscale',\n",
        "      shuffle = True\n",
        "      )\n",
        "\n",
        "\n",
        "    # one hot encoding\n",
        "\n",
        "    train_dataset = train_dataset.map(preprocess_data)\n",
        "\n",
        "    test_dataset = test_dataset.map(preprocess_data)\n",
        "\n",
        "    return train_dataset, test_dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model"
      ],
      "metadata": {
        "id": "4hfSBXIPNul-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class FedAvg(tf.keras.Model):\n",
        "\n",
        "    def __init__(self, num_classes=4, random_seed=42):\n",
        "        super(FedAvg, self).__init__()\n",
        "\n",
        "\n",
        "        # Layers\n",
        "        self.glorot_initializer = GlorotUniform(seed=random_seed)\n",
        "        self.rescale_layer = Rescaling(1.0 / 255)\n",
        "        self.conv1 = Conv2D(\n",
        "            16,\n",
        "            3,\n",
        "            padding=\"same\",\n",
        "            activation=\"relu\",\n",
        "            kernel_initializer=self.glorot_initializer)\n",
        "        self.pool = MaxPooling2D((2, 2))\n",
        "        self.conv2 = Conv2D(32, 3, padding=\"same\", activation=\"relu\", kernel_initializer=self.glorot_initializer)\n",
        "        self.conv3 = Conv2D(\n",
        "            64,\n",
        "            3,\n",
        "            padding=\"same\",\n",
        "            activation=\"relu\",\n",
        "            kernel_initializer=self.glorot_initializer,\n",
        "        )\n",
        "        self.flatten_layer = Flatten()\n",
        "        self.dense1 = Dense(\n",
        "            128,\n",
        "            activation=\"relu\",\n",
        "            kernel_initializer=self.glorot_initializer,\n",
        "            kernel_constraint=MaxNorm(3),\n",
        "        )\n",
        "        self.dropout_layer = Dropout(0.3, seed=random_seed)\n",
        "        self.dense2 = Dense(\n",
        "            num_classes,\n",
        "            activation=\"softmax\",\n",
        "            kernel_initializer=self.glorot_initializer,\n",
        "        )\n",
        "\n",
        "    def call(self, inputs):\n",
        "\n",
        "        x = self.rescale_layer(inputs)\n",
        "        x = self.conv1(x)\n",
        "        x = self.pool(x)\n",
        "        x = self.conv2(x)\n",
        "        x = self.pool(x)\n",
        "        x = self.conv3(x)\n",
        "        x = self.pool(x)\n",
        "\n",
        "        x = self.flatten_layer(x)\n",
        "        x = self.dense1(x)\n",
        "        x = self.dropout_layer(x)\n",
        "        x = self.dense2(x)\n",
        "\n",
        "        return x\n",
        "\n",
        "    \"\"\"\n",
        "    @tf.function\n",
        "    def train_step(self, inputs, labels):\n",
        "        with tf.GradientTape() as tape:\n",
        "            predictions = self(inputs, training=True)\n",
        "            loss = tf.keras.losses.categorical_crossentropy(labels, predictions)\n",
        "\n",
        "        gradients = tape.gradient(loss, self.trainable_variables)\n",
        "        self.optimizer.apply_gradients(zip(gradients, self.trainable_variables))\n",
        "        return loss\n",
        "\n",
        "    def val_step(self, x_val, y_val):\n",
        "        predictions = self(x_val, training=False)  # Predict using the model\n",
        "\n",
        "        # Calculate loss (you can use a different metric if needed)\n",
        "        loss = tf.keras.losses.categorical_crossentropy(y_val, predictions)\n",
        "\n",
        "        return loss\n",
        "    \"\"\"\n",
        "\n"
      ],
      "metadata": {
        "id": "3sLr4HWwNs3M"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Brain Tumor"
      ],
      "metadata": {
        "id": "Sj0nStyxNn2_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Dataset Creation and exploration"
      ],
      "metadata": {
        "id": "q52KoxchNQet"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset, test_dataset = create_dataset(train_path, test_path,BRAIN_TUMOR)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QUPROz5e9jo9",
        "outputId": "5389b2fe-58ab-4dc0-836c-37a77e1d1ea0"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1281 files belonging to 4 classes.\n",
            "Found 5505 files belonging to 4 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_count = count_classes(train_dataset)\n",
        "\n",
        "test_count = count_classes(test_dataset)\n",
        "\n",
        "fig, axes = plt.subplots(1, 2, figsize=(21, 5))\n",
        "\n",
        "class_names = LABELS_ALZ if train_path == DATASET_TRAIN_PATH_ALZ else LABELS_TUMOR\n",
        "\n",
        "# Plot for training set\n",
        "ax1 = axes[0]\n",
        "ax1.bar(class_names, train_count, edgecolor=\"black\", color=\"skyblue\", width=0.7)\n",
        "ax1.set_title(\"Train Distribution\")\n",
        "\n",
        "# Plot for test set\n",
        "ax3 = axes[1]\n",
        "ax3.bar(class_names, test_count, edgecolor=\"black\", color=\"lightcoral\", width=0.7)\n",
        "ax3.set_title(\"TestDistribution\")\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "XBfAzddRLLl7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "outputId": "ed56770b-77a2-4543-cc2b-a9f90e884fc4"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2100x500 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cF1tgvtsumDO"
      },
      "source": [
        "## Training"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def history_plot(history):\n",
        "  plt.plot(history.history['loss'], label='Train Loss')\n",
        "  plt.plot(history.history['val_loss'], label='Validation Loss')\n",
        "  plt.title('Training and Validation Loss')\n",
        "  plt.xlabel('Epoch')\n",
        "  plt.ylabel('Loss')\n",
        "  plt.legend()\n",
        "  plt.show()"
      ],
      "metadata": {
        "id": "Q71pJzwnbytJ"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def predictions(dataset, model):\n",
        "  y_true = []\n",
        "  y_pred = []\n",
        "\n",
        "  for images, labels in dataset:\n",
        "      y_true.extend(tf.argmax(labels, axis=1).numpy())\n",
        "      predictions = model(images, training=False)\n",
        "\n",
        "      y_pred.extend(tf.argmax(predictions, axis=1).numpy())\n",
        "  return y_true, y_pred"
      ],
      "metadata": {
        "id": "l7Ad3gPdXV-1"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define the ModelCheckpoint callback\n",
        "checkpoint_filepath = '/content/drive/MyDrive/blockchain/best_model_tum_test'\n",
        "model_checkpoint = ModelCheckpoint(\n",
        "    checkpoint_filepath,\n",
        "    save_best_only=True,\n",
        "    monitor='val_f1_score',  # Monitor the validation macro-F1 score\n",
        "    mode='max',\n",
        "    verbose=0,\n",
        "    save_weights_only=False  # Save the entire model\n",
        ")\n",
        "\n",
        "# Create and compile the model\n",
        "model = FedAvg(num_classes=4)\n",
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=[tfa.metrics.F1Score(num_classes=4,average='macro')])\n",
        "\n",
        "# Train the model\n",
        "epochs = 80\n",
        "history = model.fit(train_dataset, epochs=epochs, validation_data=test_dataset, callbacks=[model_checkpoint])\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t6OhPvL5-stC",
        "outputId": "dd5de17a-1f84-4cb4-e29b-cd6b0fb2f6c9"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/80\n",
            "173/173 [==============================] - 14s 42ms/step - loss: 0.6307 - f1_score: 0.7368 - val_loss: 0.7171 - val_f1_score: 0.6883\n",
            "Epoch 2/80\n",
            "173/173 [==============================] - 7s 42ms/step - loss: 0.3702 - f1_score: 0.8555 - val_loss: 0.3920 - val_f1_score: 0.8314\n",
            "Epoch 3/80\n",
            "173/173 [==============================] - 7s 38ms/step - loss: 0.2659 - f1_score: 0.9007 - val_loss: 0.2680 - val_f1_score: 0.8851\n",
            "Epoch 4/80\n",
            "173/173 [==============================] - 7s 38ms/step - loss: 0.2338 - f1_score: 0.9105 - val_loss: 0.2497 - val_f1_score: 0.8906\n",
            "Epoch 5/80\n",
            "173/173 [==============================] - 8s 47ms/step - loss: 0.1728 - f1_score: 0.9340 - val_loss: 0.2070 - val_f1_score: 0.9168\n",
            "Epoch 6/80\n",
            "173/173 [==============================] - 6s 34ms/step - loss: 0.1579 - f1_score: 0.9435 - val_loss: 0.2027 - val_f1_score: 0.9193\n",
            "Epoch 7/80\n",
            "173/173 [==============================] - 8s 47ms/step - loss: 0.1408 - f1_score: 0.9517 - val_loss: 0.1862 - val_f1_score: 0.9280\n",
            "Epoch 8/80\n",
            "173/173 [==============================] - 7s 38ms/step - loss: 0.1128 - f1_score: 0.9593 - val_loss: 0.1539 - val_f1_score: 0.9393\n",
            "Epoch 9/80\n",
            "173/173 [==============================] - 6s 37ms/step - loss: 0.0957 - f1_score: 0.9646 - val_loss: 0.1757 - val_f1_score: 0.9315\n",
            "Epoch 10/80\n",
            "173/173 [==============================] - 6s 34ms/step - loss: 0.0809 - f1_score: 0.9719 - val_loss: 0.1341 - val_f1_score: 0.9495\n",
            "Epoch 11/80\n",
            "173/173 [==============================] - 6s 34ms/step - loss: 0.0658 - f1_score: 0.9782 - val_loss: 0.1586 - val_f1_score: 0.9451\n",
            "Epoch 12/80\n",
            "173/173 [==============================] - 7s 42ms/step - loss: 0.0692 - f1_score: 0.9755 - val_loss: 0.1227 - val_f1_score: 0.9560\n",
            "Epoch 13/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0505 - f1_score: 0.9819 - val_loss: 0.1385 - val_f1_score: 0.9486\n",
            "Epoch 14/80\n",
            "173/173 [==============================] - 8s 44ms/step - loss: 0.0498 - f1_score: 0.9821 - val_loss: 0.1008 - val_f1_score: 0.9630\n",
            "Epoch 15/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0491 - f1_score: 0.9840 - val_loss: 0.1192 - val_f1_score: 0.9609\n",
            "Epoch 16/80\n",
            "173/173 [==============================] - 7s 39ms/step - loss: 0.0498 - f1_score: 0.9832 - val_loss: 0.2579 - val_f1_score: 0.9330\n",
            "Epoch 17/80\n",
            "173/173 [==============================] - 6s 34ms/step - loss: 0.0587 - f1_score: 0.9809 - val_loss: 0.1133 - val_f1_score: 0.9649\n",
            "Epoch 18/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0489 - f1_score: 0.9842 - val_loss: 0.1695 - val_f1_score: 0.9458\n",
            "Epoch 19/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0573 - f1_score: 0.9814 - val_loss: 0.1214 - val_f1_score: 0.9631\n",
            "Epoch 20/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0428 - f1_score: 0.9872 - val_loss: 0.1174 - val_f1_score: 0.9630\n",
            "Epoch 21/80\n",
            "173/173 [==============================] - 6s 35ms/step - loss: 0.0348 - f1_score: 0.9883 - val_loss: 0.1241 - val_f1_score: 0.9722\n",
            "Epoch 22/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0271 - f1_score: 0.9906 - val_loss: 0.1140 - val_f1_score: 0.9649\n",
            "Epoch 23/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0233 - f1_score: 0.9926 - val_loss: 0.1074 - val_f1_score: 0.9645\n",
            "Epoch 24/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0231 - f1_score: 0.9926 - val_loss: 0.0999 - val_f1_score: 0.9689\n",
            "Epoch 25/80\n",
            "173/173 [==============================] - 5s 27ms/step - loss: 0.0239 - f1_score: 0.9936 - val_loss: 0.1464 - val_f1_score: 0.9620\n",
            "Epoch 26/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0233 - f1_score: 0.9921 - val_loss: 0.1813 - val_f1_score: 0.9536\n",
            "Epoch 27/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0430 - f1_score: 0.9849 - val_loss: 0.1589 - val_f1_score: 0.9497\n",
            "Epoch 28/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0368 - f1_score: 0.9897 - val_loss: 0.1252 - val_f1_score: 0.9656\n",
            "Epoch 29/80\n",
            "173/173 [==============================] - 7s 39ms/step - loss: 0.0347 - f1_score: 0.9883 - val_loss: 0.1294 - val_f1_score: 0.9663\n",
            "Epoch 30/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0340 - f1_score: 0.9894 - val_loss: 0.1577 - val_f1_score: 0.9594\n",
            "Epoch 31/80\n",
            "173/173 [==============================] - 5s 31ms/step - loss: 0.0363 - f1_score: 0.9883 - val_loss: 0.1846 - val_f1_score: 0.9428\n",
            "Epoch 32/80\n",
            "173/173 [==============================] - 6s 33ms/step - loss: 0.0263 - f1_score: 0.9910 - val_loss: 0.1838 - val_f1_score: 0.9484\n",
            "Epoch 33/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0298 - f1_score: 0.9891 - val_loss: 0.1862 - val_f1_score: 0.9532\n",
            "Epoch 34/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0219 - f1_score: 0.9941 - val_loss: 0.1245 - val_f1_score: 0.9646\n",
            "Epoch 35/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0221 - f1_score: 0.9943 - val_loss: 0.1549 - val_f1_score: 0.9575\n",
            "Epoch 36/80\n",
            "173/173 [==============================] - 7s 37ms/step - loss: 0.0404 - f1_score: 0.9881 - val_loss: 0.1871 - val_f1_score: 0.9550\n",
            "Epoch 37/80\n",
            "173/173 [==============================] - 5s 27ms/step - loss: 0.0270 - f1_score: 0.9917 - val_loss: 0.2501 - val_f1_score: 0.9366\n",
            "Epoch 38/80\n",
            "173/173 [==============================] - 6s 33ms/step - loss: 0.0289 - f1_score: 0.9917 - val_loss: 0.1243 - val_f1_score: 0.9655\n",
            "Epoch 39/80\n",
            "173/173 [==============================] - 6s 31ms/step - loss: 0.0292 - f1_score: 0.9910 - val_loss: 0.2095 - val_f1_score: 0.9531\n",
            "Epoch 40/80\n",
            "173/173 [==============================] - 5s 27ms/step - loss: 0.0225 - f1_score: 0.9916 - val_loss: 0.1613 - val_f1_score: 0.9527\n",
            "Epoch 41/80\n",
            "173/173 [==============================] - 7s 37ms/step - loss: 0.0272 - f1_score: 0.9914 - val_loss: 0.1461 - val_f1_score: 0.9548\n",
            "Epoch 42/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0217 - f1_score: 0.9932 - val_loss: 0.1274 - val_f1_score: 0.9687\n",
            "Epoch 43/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0107 - f1_score: 0.9959 - val_loss: 0.1765 - val_f1_score: 0.9554\n",
            "Epoch 44/80\n",
            "173/173 [==============================] - 6s 37ms/step - loss: 0.0137 - f1_score: 0.9958 - val_loss: 0.1285 - val_f1_score: 0.9662\n",
            "Epoch 45/80\n",
            "173/173 [==============================] - 7s 40ms/step - loss: 0.0089 - f1_score: 0.9963 - val_loss: 0.1557 - val_f1_score: 0.9560\n",
            "Epoch 46/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0157 - f1_score: 0.9943 - val_loss: 0.1444 - val_f1_score: 0.9689\n",
            "Epoch 47/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0034 - f1_score: 0.9994 - val_loss: 0.1509 - val_f1_score: 0.9694\n",
            "Epoch 48/80\n",
            "173/173 [==============================] - 5s 27ms/step - loss: 0.0226 - f1_score: 0.9942 - val_loss: 0.1324 - val_f1_score: 0.9586\n",
            "Epoch 49/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0360 - f1_score: 0.9874 - val_loss: 0.1981 - val_f1_score: 0.9458\n",
            "Epoch 50/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0499 - f1_score: 0.9854 - val_loss: 0.1578 - val_f1_score: 0.9547\n",
            "Epoch 51/80\n",
            "173/173 [==============================] - 7s 39ms/step - loss: 0.0193 - f1_score: 0.9938 - val_loss: 0.1316 - val_f1_score: 0.9610\n",
            "Epoch 52/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0159 - f1_score: 0.9951 - val_loss: 0.1152 - val_f1_score: 0.9677\n",
            "Epoch 53/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0110 - f1_score: 0.9963 - val_loss: 0.1193 - val_f1_score: 0.9646\n",
            "Epoch 54/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0243 - f1_score: 0.9916 - val_loss: 0.1358 - val_f1_score: 0.9556\n",
            "Epoch 55/80\n",
            "173/173 [==============================] - 6s 36ms/step - loss: 0.0134 - f1_score: 0.9948 - val_loss: 0.1895 - val_f1_score: 0.9602\n",
            "Epoch 56/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0120 - f1_score: 0.9972 - val_loss: 0.1692 - val_f1_score: 0.9589\n",
            "Epoch 57/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0174 - f1_score: 0.9962 - val_loss: 0.1626 - val_f1_score: 0.9598\n",
            "Epoch 58/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0119 - f1_score: 0.9967 - val_loss: 0.1492 - val_f1_score: 0.9602\n",
            "Epoch 59/80\n",
            "173/173 [==============================] - 7s 38ms/step - loss: 0.0093 - f1_score: 0.9974 - val_loss: 0.1798 - val_f1_score: 0.9571\n",
            "Epoch 60/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0232 - f1_score: 0.9916 - val_loss: 0.2067 - val_f1_score: 0.9509\n",
            "Epoch 61/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0280 - f1_score: 0.9905 - val_loss: 0.1780 - val_f1_score: 0.9591\n",
            "Epoch 62/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0366 - f1_score: 0.9868 - val_loss: 0.2006 - val_f1_score: 0.9480\n",
            "Epoch 63/80\n",
            "173/173 [==============================] - 7s 37ms/step - loss: 0.0277 - f1_score: 0.9914 - val_loss: 0.1637 - val_f1_score: 0.9611\n",
            "Epoch 64/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0245 - f1_score: 0.9922 - val_loss: 0.2306 - val_f1_score: 0.9569\n",
            "Epoch 65/80\n",
            "173/173 [==============================] - 6s 37ms/step - loss: 0.0073 - f1_score: 0.9984 - val_loss: 0.1488 - val_f1_score: 0.9670\n",
            "Epoch 66/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0105 - f1_score: 0.9972 - val_loss: 0.1373 - val_f1_score: 0.9687\n",
            "Epoch 67/80\n",
            "173/173 [==============================] - 7s 40ms/step - loss: 0.0017 - f1_score: 1.0000 - val_loss: 0.1162 - val_f1_score: 0.9737\n",
            "Epoch 68/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0034 - f1_score: 0.9985 - val_loss: 0.1780 - val_f1_score: 0.9654\n",
            "Epoch 69/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0041 - f1_score: 0.9985 - val_loss: 0.1573 - val_f1_score: 0.9689\n",
            "Epoch 70/80\n",
            "173/173 [==============================] - 7s 38ms/step - loss: 0.0095 - f1_score: 0.9967 - val_loss: 0.1845 - val_f1_score: 0.9655\n",
            "Epoch 71/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0081 - f1_score: 0.9974 - val_loss: 0.2470 - val_f1_score: 0.9547\n",
            "Epoch 72/80\n",
            "173/173 [==============================] - 6s 33ms/step - loss: 0.0256 - f1_score: 0.9917 - val_loss: 0.2127 - val_f1_score: 0.9540\n",
            "Epoch 73/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0183 - f1_score: 0.9932 - val_loss: 0.1933 - val_f1_score: 0.9609\n",
            "Epoch 74/80\n",
            "173/173 [==============================] - 7s 39ms/step - loss: 0.0260 - f1_score: 0.9923 - val_loss: 0.1172 - val_f1_score: 0.9626\n",
            "Epoch 75/80\n",
            "173/173 [==============================] - 5s 27ms/step - loss: 0.0191 - f1_score: 0.9936 - val_loss: 0.1641 - val_f1_score: 0.9600\n",
            "Epoch 76/80\n",
            "173/173 [==============================] - 6s 32ms/step - loss: 0.0227 - f1_score: 0.9912 - val_loss: 0.2074 - val_f1_score: 0.9493\n",
            "Epoch 77/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0383 - f1_score: 0.9893 - val_loss: 0.1908 - val_f1_score: 0.9526\n",
            "Epoch 78/80\n",
            "173/173 [==============================] - 8s 46ms/step - loss: 0.0301 - f1_score: 0.9882 - val_loss: 0.1561 - val_f1_score: 0.9649\n",
            "Epoch 79/80\n",
            "173/173 [==============================] - 5s 28ms/step - loss: 0.0133 - f1_score: 0.9958 - val_loss: 0.1657 - val_f1_score: 0.9599\n",
            "Epoch 80/80\n",
            "173/173 [==============================] - 6s 35ms/step - loss: 0.0065 - f1_score: 0.9978 - val_loss: 0.1520 - val_f1_score: 0.9578\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history_plot(history)"
      ],
      "metadata": {
        "id": "i5WESdCDcERg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "101dabd3-567c-4257-9012-6518b389dd7e"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"The best epoch is: {model_checkpoint.best:.3f}\")\n",
        "\n",
        "# Load the model with the best weights\n",
        "best_model_path = '/content/drive/MyDrive/blockchain/best_model_tum_test'\n",
        "best_model = load_model(best_model_path)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JrsmkVuRbOgm",
        "outputId": "cf5f5eaf-c8fb-411c-c955-c267ef972a13"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The best epoch is: 0.974\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_true, y_pred = predictions(test_dataset, best_model)\n",
        "# Generate and print the classification report\n",
        "class_names = LABELS_ALZ if train_path == DATASET_TRAIN_PATH_ALZ else LABELS_TUMOR\n",
        "\n",
        "print(\"\\t\\tClassification report:\\n\")\n",
        "print(classification_report(y_true, y_pred,target_names=class_names))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "57aT2r6v-eZF",
        "outputId": "255fad1b-cf88-41cf-b55f-21242cd3a03e"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\t\tClassification report:\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "      glioma       0.98      0.95      0.96       299\n",
            "  meningioma       0.94      0.96      0.95       301\n",
            "     notumor       0.99      1.00      0.99       381\n",
            "   pituitary       0.99      0.98      0.98       300\n",
            "\n",
            "    accuracy                           0.98      1281\n",
            "   macro avg       0.97      0.97      0.97      1281\n",
            "weighted avg       0.98      0.98      0.98      1281\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Alzheimer"
      ],
      "metadata": {
        "id": "qGtgJGDkDeVH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_path = DATASET_TRAIN_PATH_ALZ\n",
        "test_path = DATASET_TEST_PATH_ALZ\n",
        "set_riproducibility()"
      ],
      "metadata": {
        "id": "sk-EB1H7DibR"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset, test_dataset = create_dataset(train_path, test_path,ALZHEIMER)"
      ],
      "metadata": {
        "id": "RpOBFJQGDhf8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d721835a-b622-4183-8eec-6b427dc4c4f3"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1025 files belonging to 4 classes.\n",
            "Found 4096 files belonging to 4 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_count = count_classes(train_dataset)\n",
        "test_count = count_classes(test_dataset)\n",
        "\n",
        "fig, axes = plt.subplots(1, 2, figsize=(21, 5))\n",
        "\n",
        "class_names = LABELS_ALZ if train_path == DATASET_TRAIN_PATH_ALZ else LABELS_TUMOR\n",
        "\n",
        "# Plot for training set\n",
        "ax1 = axes[0]\n",
        "ax1.bar(class_names, train_count, edgecolor=\"black\", color=\"skyblue\", width=0.7)\n",
        "ax1.set_title(\"Train Distribution\")\n",
        "\n",
        "# Plot for test set\n",
        "ax3 = axes[1]\n",
        "ax3.bar(class_names, test_count, edgecolor=\"black\", color=\"lightcoral\", width=0.7)\n",
        "ax3.set_title(\"TestDistribution\")\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "id": "UDAuNq_9DwiR",
        "outputId": "0fbc96fc-2f3f-4010-cc2c-e0ca6cec849c"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2100x500 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABpwAAAHDCAYAAAA5ormYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABj/klEQVR4nO39e5xWZb0//r+GwwzHAQFhIDlpHkAFPBRSHlAJVHJr2q484hZF3aglZW7abkOsMG2rWahZClrw0WyrlZqKlocUTxSeIzUMSwbCghFRDrJ+f/Tj/joBCtwzDujz+Xisx4P7uq57rWvNve7l/fZ132tVFEVRBAAAAAAAADZRs6aeAAAAAAAAAFs2gRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETwIfECSeckD59+jTpHKZOnZqKioq8/PLLjb6tf93fl19+ORUVFfnOd77T6NtOkgkTJqSiouJ92RYAAEBjeT9ryT59+uSEE04oPV5TQz7xxBPvy/aHDh2aoUOHvi/bAvggEjgBNLGKiooNWu67776mnmo99913X735VVVVpVu3bhk6dGi+9a1v5W9/+1uDbGfZsmWZMGHCZrf/yeY9NwAAYPPxftZ971anrPli3JqlTZs26dWrVw499NBMmTIly5cvL3v7SfLcc89lwoQJ78uXDTfW5jw3gC1di6aeAMCH3Y9//ON6j6+//vrMmDFjrfZ+/fqVtZ0f/vCHWb16dVnrWJczzzwzH/vYx/L222/nb3/7Wx5++OF8/etfzyWXXJKf/vSnOeCAA0pjjzvuuHzhC19IVVXVBq9/2bJlOf/885Nko75p1lj7+07vNrdzzz03//Vf/9Wo2wcAALYM71fdl2xYDXXllVemXbt2Wb58ef7617/mrrvuyoknnpjLLrsst912W3r27Fkauym11XPPPZfzzz8/Q4cO3ahfR82ZMyfNmjXu9+PfbW533313o24b4INO4ATQxI499th6jx955JHMmDFjrfZ/tWzZsrRp02aDt9OyZctNmt972WefffLZz362XtuTTz6Z4cOH58gjj8xzzz2X7t27J0maN2+e5s2bN8o81njjjTfStm3bRtvfDdWiRYu0aOE/swAAwKbXfY3ls5/9bLp06VJ6fN5552XatGk5/vjj8+///u955JFHSn2NXVsVRZG33norrVu33qgvJzaGysrKJt0+wJbOJfUAtgBDhw7NLrvsklmzZmXfffdNmzZt8rWvfS1J8vOf/zwjR45Mjx49UlVVle222y4XXHBB3n777XrreLd7Gl199dXZbrvtUlVVlY997GN5/PHHy5rvwIEDc9lll2Xx4sX5/ve/X2pf1z2cnnjiiYwYMSJdunRJ69at07dv35x44omlOW699dZJkvPPP7902YcJEyaU9qldu3Z56aWXcsghh6R9+/Y55phj1rm/73TppZemd+/ead26dfbbb78888wz9frXd93ud67zvea2rns4rVq1KhdccEHpb92nT5987WtfW+uyFX369MmnP/3p/Pa3v83HP/7xtGrVKttuu22uv/76df/BAQCALd7q1atz2WWXZeedd06rVq3SrVu3nHLKKfnHP/5Rb1w5NdS7OeaYY3LSSSfl0UcfzYwZM0rt66qtbrjhhuyxxx5p3759qqurs+uuu+a73/1ukn/Wff/+7/+eJNl///3Xulzgmnrnrrvuyp577pnWrVvnBz/4QanvnfdwWmPZsmU55ZRT0rlz51RXV+f4449f6++yvv185zrfa27rqgUXLlyY0aNHp1u3bmnVqlUGDhyY6667rt6YxqyvAbYkvnoNsIV47bXXcvDBB+cLX/hCjj322HTr1i3JPz8wt2vXLuPGjUu7du3y61//Ouedd17q6upy8cUXv+d6p0+fntdffz2nnHJKKioqctFFF+WII47In/70p7K+yfbZz342o0ePzt13351vfvOb6xyzcOHCDB8+PFtvvXX+67/+Kx07dszLL7+cm2++OUmy9dZb58orr8xpp52Wz3zmMzniiCOSJAMGDCitY9WqVRkxYkT23nvvfOc733nPX31df/31ef311zN27Ni89dZb+e53v5sDDjggTz/9dOlvuiE2ZG7/6qSTTsp1112Xz372s/nyl7+cRx99NJMmTcrzzz+fW265pd7YF198sfQ3HDVqVK699tqccMIJ2WOPPbLzzjtv8DwBAIAtwymnnJKpU6fmP/7jP3LmmWdm7ty5+f73v5/f//73eeihh9KyZcsGqaHezXHHHZerr746d999dz71qU+tc8yMGTNy1FFH5cADD8y3v/3tJMnzzz+fhx56KF/84hez77775swzz8zll1+er33ta6XLBL7zcoFz5szJUUcdlVNOOSUnn3xydtxxx3ed1+mnn56OHTtmwoQJmTNnTq688sr8+c9/Lt1beENtyNze6c0338zQoUPz4osv5vTTT0/fvn1z00035YQTTsjixYvzxS9+sd74xqqvAbYUAieALURtbW2uuuqqnHLKKfXap0+fntatW5cen3rqqTn11FNzxRVX5Bvf+MZ7XpJg3rx5eeGFF7LVVlslSXbcccccdthhueuuu/LpT396k+fbsmXL7LDDDnnppZfWO+bhhx/OP/7xj9x9993Zc889S+3f+MY3kiRt27bNZz/72Zx22mkZMGDAOi83sXz58vz7v/97Jk2atEHzevHFF/PCCy/kIx/5SJLkoIMOyuDBg/Ptb387l1xyyQbv34bM7Z2efPLJXHfddTnppJPywx/+MEnyn//5n+natWu+853v5De/+U3233//0vg5c+bkgQceyD777JMk+dznPpeePXtmypQp+c53vrPB8wQAADZ/v/3tb/OjH/0o06ZNy9FHH11q33///XPQQQflpptuytFHH90gNdS72WWXXZLkXeu422+/PdXV1bnrrrvWecn0bbfdNvvss08uv/zyfOpTn1rn1SNefPHF3HnnnRkxYsQGzauysjL33ntvKbTp3bt3vvrVr+aXv/xl/u3f/m2D1rGhc3unq6++Os8//3x+8pOflK6mceqpp2a//fbLueeemxNPPDHt27cvjW+s+hpgS+GSegBbiKqqqvzHf/zHWu3vDJtef/31LFq0KPvss0+WLVuWP/zhD++53s9//vOlD8NJSgHHn/70p7Ln3K5du7z++uvr7e/YsWOS5LbbbsvKlSs3eTunnXbaBo89/PDDS2FTknz84x/P4MGDc8cdd2zy9jfEmvWPGzeuXvuXv/zlJP8s2t6pf//+pdci+ec3FXfccccGeV0AAIDNy0033ZQOHTrkU5/6VBYtWlRa9thjj7Rr1y6/+c1vkjRcDbU+7dq1S5L3rOPeeOONepfd21h9+/bd4LApScaMGVPvF0KnnXZaWrRo8b7UcTU1NTnqqKNKbS1btsyZZ56ZpUuX5v777683vjHra4AtgcAJYAvxkY98ZJ03MH322Wfzmc98Jh06dEh1dXW23nrr0rfYlixZ8p7r7dWrV73Haz4c/+v1sDfF0qVL633b61/tt99+OfLII3P++eenS5cuOeywwzJlypS17mn0blq0aJFtttlmg8dvv/32a7XtsMMO9e4r1Rj+/Oc/p1mzZvnoRz9ar72mpiYdO3bMn//853rt//q6JP98bRridQEAADYvL7zwQpYsWZKuXbtm6623rrcsXbo0CxcuTNIwNdS7Wbp0aZK8ax33n//5n9lhhx1y8MEHZ5tttsmJJ56YO++8c6O207dv340a/691XLt27dK9e/f3pY7bfvvt06xZ/f+FuuYSfO9VxzVkfQ2wJXBJPYAtxDt/ybTG4sWLs99++6W6ujoTJ07Mdtttl1atWuV3v/tdzjnnnKxevfo917uuSyAkSVEUZc135cqV+eMf/1i6JMO6VFRU5Gc/+1keeeSR/PKXv8xdd92VE088Mf/7v/+bRx55pPTtundTVVW11of/clVUVKxz/99+++0GWfeGaKzXBQAA2PysXr06Xbt2zbRp09bZv/XWWydpmBrq3TzzzDNJstYX5d6pa9eumT17du6666786le/yq9+9atMmTIlxx9/fK677roN2s666tvG0hB13IZSxwEfdn7hBLAFu++++/Laa69l6tSp+eIXv5hPf/rTGTZsWL2f8DeVn/3sZ3nzzTc36DIJe+21V775zW/miSeeyLRp0/Lss8/mhhtuSLLhAc2GeuGFF9Zq++Mf/5g+ffqUHm+11VZZvHjxWuP+9dtrGzO33r17Z/Xq1Wttf8GCBVm8eHF69+69wesCAAA+WLbbbru89tpr+eQnP5lhw4attQwcOLDe+MaqoX784x8nyXvWcZWVlTn00ENzxRVX5KWXXsopp5yS66+/Pi+++GLZc1iXf62jli5dmvnz579nHbdixYrMnz+/XtvG1nEvvPDCWl/mXHP5enUcQH0CJ4At2JpvT73z21IrVqzIFVdc0VRTSpI8+eST+dKXvpStttoqY8eOXe+4f/zjH2t902vQoEFJUrokRJs2bZJknQHQprj11lvz17/+tfT4sccey6OPPpqDDz641LbddtvlD3/4Q/72t7+V2p588sk89NBD9da1MXM75JBDkiSXXXZZvfZLLrkkSTJy5MiN2g8AAOCD43Of+1zefvvtXHDBBWv1rVq1qlRzNGYNNX369PzoRz/KkCFDcuCBB6533GuvvVbvcbNmzTJgwIB6c2jbtu0mzWF9rr766nr3rLryyiuzatWqteq4Bx54YK3n/esvnDZmboccckhqa2tz4403ltpWrVqV733ve2nXrl3222+/TdkdgA8sl9QD2IJ94hOfyFZbbZVRo0blzDPPTEVFRX784x+/rz/Xf/DBB/PWW2/l7bffzmuvvZaHHnoov/jFL9KhQ4fccsstqampWe9zr7vuulxxxRX5zGc+k+222y6vv/56fvjDH6a6uroU0LRu3Tr9+/fPjTfemB122CGdOnXKLrvs8q6X6ns3H/3oR7P33nvntNNOy/Lly3PZZZelc+fO+epXv1oac+KJJ+aSSy7JiBEjMnr06CxcuDBXXXVVdt5559TV1ZXGbczcBg4cmFGjRuXqq68uXQrxsccey3XXXZfDDz88+++//ybtDwAAsOXbb7/9csopp2TSpEmZPXt2hg8fnpYtW+aFF17ITTfdlO9+97v57Gc/22A11M9+9rO0a9cuK1asyF//+tfcddddeeihhzJw4MDcdNNN7zrXk046KX//+99zwAEHZJtttsmf//znfO9738ugQYNK9zYaNGhQmjdvnm9/+9tZsmRJqqqqcsABB6Rr166b9PdZsWJFDjzwwHzuc5/LnDlzcsUVV2TvvffOv/3bv9Wb16mnnpojjzwyn/rUp/Lkk0/mrrvuSpcuXeqta2PmNmbMmPzgBz/ICSeckFmzZqVPnz752c9+loceeiiXXXbZu97rCuDDSOAEsAXr3Llzbrvttnz5y1/Oueeem6222irHHntsDjzwwA26lF1DuPzyy5MkLVu2TMeOHdOvX7+cf/75Ofnkk0vXGV+fNaHLDTfckAULFqRDhw75+Mc/nmnTptW7ieyPfvSjnHHGGTnrrLOyYsWKfP3rX9/kwOn4449Ps2bNctlll2XhwoX5+Mc/nu9///vp3r17aUy/fv1y/fXX57zzzsu4cePSv3///PjHP8706dNz33331VvfxsztRz/6UbbddttMnTq1FMaNHz8+X//61zdpXwAAgA+Oq666KnvssUd+8IMf5Gtf+1patGiRPn365Nhjj80nP/nJJA1XQ5122mlJklatWqVLly4ZNGhQrr322hx99NGpqqp613kee+yxufrqq3PFFVdk8eLFqampyec///lMmDChdH/dmpqaXHXVVZk0aVJGjx6dt99+O7/5zW82OXD6/ve/n2nTpuW8887LypUrc9RRR+Xyyy+vd3m8k08+OXPnzs0111yTO++8M/vss09mzJix1q+1NmZurVu3zn333Zf/+q//ynXXXZe6urrsuOOOmTJlSk444YRN2heAD7KKwl3rAAAAAAAAKIN7OAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGVp0dQTaCyrV6/Oq6++mvbt26eioqKppwMAAB8qRVHk9ddfT48ePdKsme+5UR71HQAANI2Nqe0+sIHTq6++mp49ezb1NAAA4EPtlVdeyTbbbNPU02ALp74DAICmtSG13Qc2cGrfvn2Sf/4Rqqurm3g2AADw4VJXV5eePXuWPpdDOdR3AADQNDamtvvABk5rLrNQXV2tIAEAgCbi8mc0BPUdAAA0rQ2p7VxMHQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICytGjqCXzQzZs3L4sWLWrqadBEunTpkl69ejX1NAAAgAagvvvwUtsBALw3gVMjmjdvXnbq1y9vLlvW1FOhibRu0yZ/eP55hQkAAGzh5s2bl3477ZRlb77Z1FOhCbRp3TrP/+EPajsAgHchcGpEixYtypvLluVz37gyXftu39TT4X22cO4L+em5p2XRokWKEgAA2MItWrQoy958M1cfcUR26NKlqafD++iPixZlzM03q+0AAN6DwOl90LXv9vlIv4FNPQ0AAADKtEOXLhnUo0dTTwMAADY7zZp6AgAAAAAAAGzZNipwmjRpUj72sY+lffv26dq1aw4//PDMmTOn3pi33norY8eOTefOndOuXbsceeSRWbBgQb0x8+bNy8iRI9OmTZt07do1Z599dlatWlVvzH333Zfdd989VVVV+ehHP5qpU6du2h4CAACwSSZMmJCKiop6y0477VTqb6j6DwAA2PJtVOB0//33Z+zYsXnkkUcyY8aMrFy5MsOHD88bb7xRGnPWWWfll7/8ZW666abcf//9efXVV3PEEUeU+t9+++2MHDkyK1asyMMPP5zrrrsuU6dOzXnnnVcaM3fu3IwcOTL7779/Zs+enS996Us56aSTctdddzXALgMAALChdt5558yfP7+0/Pa3vy31NUT9BwAAfDBs1D2c7rzzznqPp06dmq5du2bWrFnZd999s2TJklxzzTWZPn16DjjggCTJlClT0q9fvzzyyCPZa6+9cvfdd+e5557LPffck27dumXQoEG54IILcs4552TChAmprKzMVVddlb59++Z///d/kyT9+vXLb3/721x66aUZMWJEA+06AAAA76VFixapqalZq72h6j8AAOCDoax7OC1ZsiRJ0qlTpyTJrFmzsnLlygwbNqw0ZqeddkqvXr0yc+bMJMnMmTOz6667plu3bqUxI0aMSF1dXZ599tnSmHeuY82YNetYl+XLl6eurq7eAgAAQHleeOGF9OjRI9tuu22OOeaYzJs3L0nD1X/ror4DAIAtzyYHTqtXr86XvvSlfPKTn8wuu+ySJKmtrU1lZWU6duxYb2y3bt1SW1tbGvPOYmNN/5q+dxtTV1eXN998c53zmTRpUjp06FBaevbsuam7BgAAQJLBgwdn6tSpufPOO3PllVdm7ty52WefffL66683WP23Luo7AADY8mzUJfXeaezYsXnmmWfqXb+7KY0fPz7jxo0rPa6rq1OUAAAAlOHggw8u/XvAgAEZPHhwevfunZ/+9Kdp3bp1o21XfQcAAFueTfqF0+mnn57bbrstv/nNb7LNNtuU2mtqarJixYosXry43vgFCxaUrvldU1OTBQsWrNW/pu/dxlRXV6+3qKmqqkp1dXW9BQAAgIbTsWPH7LDDDnnxxRcbrP5bF/UdAABseTYqcCqKIqeffnpuueWW/PrXv07fvn3r9e+xxx5p2bJl7r333lLbnDlzMm/evAwZMiRJMmTIkDz99NNZuHBhacyMGTNSXV2d/v37l8a8cx1rxqxZBwAAAO+/pUuX5qWXXkr37t0brP4DAAA+GDbqknpjx47N9OnT8/Of/zzt27cvXXO7Q4cOad26dTp06JDRo0dn3Lhx6dSpU6qrq3PGGWdkyJAh2WuvvZIkw4cPT//+/XPcccfloosuSm1tbc4999yMHTs2VVVVSZJTTz013//+9/PVr341J554Yn7961/npz/9aW6//fYG3n0AAADW5ytf+UoOPfTQ9O7dO6+++mq+/vWvp3nz5jnqqKMarP4DAAA+GDYqcLryyiuTJEOHDq3XPmXKlJxwwglJkksvvTTNmjXLkUcemeXLl2fEiBG54oorSmObN2+e2267LaeddlqGDBmStm3bZtSoUZk4cWJpTN++fXP77bfnrLPOyne/+91ss802+dGPfpQRI0Zs4m4CAACwsf7yl7/kqKOOymuvvZatt946e++9dx555JFsvfXWSRqm/gMAAD4YNipwKoriPce0atUqkydPzuTJk9c7pnfv3rnjjjvedT1Dhw7N73//+42ZHgAAAA3ohhtueNf+hqr/AACALd9G3cMJAAAAAAAA/pXACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICybHTg9MADD+TQQw9Njx49UlFRkVtvvbVef0VFxTqXiy++uDSmT58+a/VfeOGF9dbz1FNPZZ999kmrVq3Ss2fPXHTRRZu2hwAAAAAAADSqjQ6c3njjjQwcODCTJ09eZ//8+fPrLddee20qKipy5JFH1hs3ceLEeuPOOOOMUl9dXV2GDx+e3r17Z9asWbn44oszYcKEXH311Rs7XQAAAAAAABpZi419wsEHH5yDDz54vf01NTX1Hv/85z/P/vvvn2233bZee/v27dcau8a0adOyYsWKXHvttamsrMzOO++c2bNn55JLLsmYMWM2dsoAAAAAAAA0oka9h9OCBQty++23Z/To0Wv1XXjhhencuXN22223XHzxxVm1alWpb+bMmdl3331TWVlZahsxYkTmzJmTf/zjH+vc1vLly1NXV1dvAQAAAAAAoPFt9C+cNsZ1112X9u3b54gjjqjXfuaZZ2b33XdPp06d8vDDD2f8+PGZP39+LrnkkiRJbW1t+vbtW+853bp1K/VttdVWa21r0qRJOf/88xtpTwAAAAAAAFifRg2crr322hxzzDFp1apVvfZx48aV/j1gwIBUVlbmlFNOyaRJk1JVVbVJ2xo/fny99dbV1aVnz56bNnEAAAAAAAA2WKMFTg8++GDmzJmTG2+88T3HDh48OKtWrcrLL7+cHXfcMTU1NVmwYEG9MWser+++T1VVVZscVgEAAAAAALDpGu0eTtdcc0322GOPDBw48D3Hzp49O82aNUvXrl2TJEOGDMkDDzyQlStXlsbMmDEjO+644zovpwcAAAAAAEDT2ejAaenSpZk9e3Zmz56dJJk7d25mz56defPmlcbU1dXlpptuykknnbTW82fOnJnLLrssTz75ZP70pz9l2rRpOeuss3LssceWwqSjjz46lZWVGT16dJ599tnceOON+e53v1vvknkAAAAAAABsHjb6knpPPPFE9t9//9LjNSHQqFGjMnXq1CTJDTfckKIoctRRR631/Kqqqtxwww2ZMGFCli9fnr59++ass86qFyZ16NAhd999d8aOHZs99tgjXbp0yXnnnZcxY8Zs7HQBAAAAAABoZBsdOA0dOjRFUbzrmDFjxqw3HNp9993zyCOPvOd2BgwYkAcffHBjpwcAAAAAAMD7rNHu4QQAAAAAAMCHg8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAgPd04YUXpqKiIl/60pdKbW+99VbGjh2bzp07p127djnyyCOzYMGCes+bN29eRo4cmTZt2qRr1645++yzs2rVqvd59gAAQGMTOAEAAPCuHn/88fzgBz/IgAED6rWfddZZ+eUvf5mbbrop999/f1599dUcccQRpf633347I0eOzIoVK/Lwww/nuuuuy9SpU3Peeee937sAAAA0MoETAAAA67V06dIcc8wx+eEPf5itttqq1L5kyZJcc801ueSSS3LAAQdkjz32yJQpU/Lwww/nkUceSZLcfffdee655/KTn/wkgwYNysEHH5wLLrggkydPzooVK5pqlwAAgEYgcAIAAGC9xo4dm5EjR2bYsGH12mfNmpWVK1fWa99pp53Sq1evzJw5M0kyc+bM7LrrrunWrVtpzIgRI1JXV5dnn332/dkBAADgfdGiqScAAADA5umGG27I7373uzz++ONr9dXW1qaysjIdO3as196tW7fU1taWxrwzbFrTv6ZvfZYvX57ly5eXHtfV1W3qLgAAAO8Tv3ACAABgLa+88kq++MUvZtq0aWnVqtX7uu1JkyalQ4cOpaVnz57v6/YBAICNJ3ACAABgLbNmzcrChQuz++67p0WLFmnRokXuv//+XH755WnRokW6deuWFStWZPHixfWet2DBgtTU1CRJampqsmDBgrX61/Stz/jx47NkyZLS8sorrzTszgEAAA1O4AQAAMBaDjzwwDz99NOZPXt2adlzzz1zzDHHlP7dsmXL3HvvvaXnzJkzJ/PmzcuQIUOSJEOGDMnTTz+dhQsXlsbMmDEj1dXV6d+//3q3XVVVlerq6noLAACweXMPJwAAANbSvn377LLLLvXa2rZtm86dO5faR48enXHjxqVTp06prq7OGWeckSFDhmSvvfZKkgwfPjz9+/fPcccdl4suuii1tbU599xzM3bs2FRVVb3v+wQAADQegRMAAACb5NJLL02zZs1y5JFHZvny5RkxYkSuuOKKUn/z5s1z22235bTTTsuQIUPStm3bjBo1KhMnTmzCWQMAAI1B4AQAAMAGue++++o9btWqVSZPnpzJkyev9zm9e/fOHXfc0cgzAwAAmpp7OAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFCWjQ6cHnjggRx66KHp0aNHKioqcuutt9brP+GEE1JRUVFvOeigg+qN+fvf/55jjjkm1dXV6dixY0aPHp2lS5fWG/PUU09ln332SatWrdKzZ89cdNFFG793AAAAAAAANLqNDpzeeOONDBw4MJMnT17vmIMOOijz588vLf/v//2/ev3HHHNMnn322cyYMSO33XZbHnjggYwZM6bUX1dXl+HDh6d3796ZNWtWLr744kyYMCFXX331xk4XAAAAAACARtZiY59w8MEH5+CDD37XMVVVVampqVln3/PPP58777wzjz/+ePbcc88kyfe+970ccsgh+c53vpMePXpk2rRpWbFiRa699tpUVlZm5513zuzZs3PJJZfUC6YAAAAAAABoeo1yD6f77rsvXbt2zY477pjTTjstr732Wqlv5syZ6dixYylsSpJhw4alWbNmefTRR0tj9t1331RWVpbGjBgxInPmzMk//vGPxpgyAAAAAAAAm2ijf+H0Xg466KAcccQR6du3b1566aV87Wtfy8EHH5yZM2emefPmqa2tTdeuXetPokWLdOrUKbW1tUmS2tra9O3bt96Ybt26lfq22mqrtba7fPnyLF++vPS4rq6uoXcNAAAAAACAdWjwwOkLX/hC6d+77rprBgwYkO222y733XdfDjzwwIbeXMmkSZNy/vnnN9r6AQAAAAAAWLdGuaTeO2277bbp0qVLXnzxxSRJTU1NFi5cWG/MqlWr8ve//71036eamposWLCg3pg1j9d3b6jx48dnyZIlpeWVV15p6F0BAAAAAABgHRo9cPrLX/6S1157Ld27d0+SDBkyJIsXL86sWbNKY379619n9erVGTx4cGnMAw88kJUrV5bGzJgxIzvuuOM6L6eXJFVVVamurq63AAAAAAAA0Pg2OnBaunRpZs+endmzZydJ5s6dm9mzZ2fevHlZunRpzj777DzyyCN5+eWXc++99+awww7LRz/60YwYMSJJ0q9fvxx00EE5+eST89hjj+Whhx7K6aefni984Qvp0aNHkuToo49OZWVlRo8enWeffTY33nhjvvvd72bcuHENt+cAAAAAAAA0iI0OnJ544onstttu2W233ZIk48aNy2677ZbzzjsvzZs3z1NPPZV/+7d/yw477JDRo0dnjz32yIMPPpiqqqrSOqZNm5addtopBx54YA455JDsvffeufrqq0v9HTp0yN133525c+dmjz32yJe//OWcd955GTNmTAPsMgAAAAAAAA2pxcY+YejQoSmKYr39d91113uuo1OnTpk+ffq7jhkwYEAefPDBjZ0eAAAAAAAA77NGv4cTAAAAAAAAH2wCJwAAAAAAAMoicAIAAAAAAKAsAicAAAAAAADK0qKpJwAAAABA05g3b14WLVrU1NOgCXTp0iW9evVq6mkA8AEicAI2moLkw0tBAgAAHxzz5s1Lv512yrI332zqqdAE2rRunef/8Ac1HgANRuAEbJR58+Zlp3798uayZU09FZpA6zZt8ofnn1eQAADAB8CiRYuy7M03c/URR2SHLl2aejq8j/64aFHG3HxzFi1apL4DoMEInICNsmjRory5bFk+940r07Xv9k09Hd5HC+e+kJ+ee5qCBAAAPmB26NIlg3r0aOppAABbOIETsEm69t0+H+k3sKmnAQAAAADAZqBZU08AAAAAAACALZvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKMtGB04PPPBADj300PTo0SMVFRW59dZbS30rV67MOeeck1133TVt27ZNjx49cvzxx+fVV1+tt44+ffqkoqKi3nLhhRfWG/PUU09ln332SatWrdKzZ89cdNFFm7aHAAAAAAAANKqNDpzeeOONDBw4MJMnT16rb9myZfnd736X//mf/8nvfve73HzzzZkzZ07+7d/+ba2xEydOzPz580vLGWecUeqrq6vL8OHD07t378yaNSsXX3xxJkyYkKuvvnpjpwsAAAAAAEAja7GxTzj44INz8MEHr7OvQ4cOmTFjRr2273//+/n4xz+eefPmpVevXqX29u3bp6amZp3rmTZtWlasWJFrr702lZWV2XnnnTN79uxccsklGTNmzMZOGQAAAAAAgEbU6PdwWrJkSSoqKtKxY8d67RdeeGE6d+6c3XbbLRdffHFWrVpV6ps5c2b23XffVFZWltpGjBiROXPm5B//+Mc6t7N8+fLU1dXVWwAAAAAAAGh8G/0Lp43x1ltv5ZxzzslRRx2V6urqUvuZZ56Z3XffPZ06dcrDDz+c8ePHZ/78+bnkkkuSJLW1tenbt2+9dXXr1q3Ut9VWW621rUmTJuX8889vxL0BAAAAAABgXRotcFq5cmU+97nPpSiKXHnllfX6xo0bV/r3gAEDUllZmVNOOSWTJk1KVVXVJm1v/Pjx9dZbV1eXnj17btrkAQAAAAAA2GCNEjitCZv+/Oc/59e//nW9Xzety+DBg7Nq1aq8/PLL2XHHHVNTU5MFCxbUG7Pm8fru+1RVVbXJYRUAAAAAAACbrsHv4bQmbHrhhRdyzz33pHPnzu/5nNmzZ6dZs2bp2rVrkmTIkCF54IEHsnLlytKYGTNmZMcdd1zn5fQAAAAAAABoOhv9C6elS5fmxRdfLD2eO3duZs+enU6dOqV79+757Gc/m9/97ne57bbb8vbbb6e2tjZJ0qlTp1RWVmbmzJl59NFHs//++6d9+/aZOXNmzjrrrBx77LGlMOnoo4/O+eefn9GjR+ecc87JM888k+9+97u59NJLG2i3AQAAAAAAaCgb/QunJ554Irvttlt22223JP+8H9Nuu+2W8847L3/961/zi1/8In/5y18yaNCgdO/evbQ8/PDDSf556bsbbrgh++23X3beeed885vfzFlnnZWrr766tI0OHTrk7rvvzty5c7PHHnvky1/+cs4777yMGTOmgXYbAACA93LllVdmwIABqa6uTnV1dYYMGZJf/epXpf633norY8eOTefOndOuXbsceeSRa10efd68eRk5cmTatGmTrl275uyzz86qVave710BAAAa2Ub/wmno0KEpimK9/e/WlyS77757HnnkkffczoABA/Lggw9u7PQAAABoINtss00uvPDCbL/99imKItddd10OO+yw/P73v8/OO++cs846K7fffntuuummdOjQIaeffnqOOOKIPPTQQ0mSt99+OyNHjkxNTU0efvjhzJ8/P8cff3xatmyZb33rW028dwAAQEPa6MAJAACAD4dDDz203uNvfvObufLKK/PII49km222yTXXXJPp06fngAMOSJJMmTIl/fr1yyOPPJK99tord999d5577rncc8896datWwYNGpQLLrgg55xzTiZMmJDKysqm2C0AAKARbPQl9QAAAPjwefvtt3PDDTfkjTfeyJAhQzJr1qysXLkyw4YNK43Zaaed0qtXr8ycOTNJMnPmzOy6667p1q1bacyIESNSV1eXZ599dr3bWr58eerq6uotAADA5k3gBAAAwHo9/fTTadeuXaqqqnLqqafmlltuSf/+/VNbW5vKysp07Nix3vhu3bqltrY2SVJbW1svbFrTv6ZvfSZNmpQOHTqUlp49ezbsTgEAAA1O4AQAAMB67bjjjpk9e3YeffTRnHbaaRk1alSee+65Rt3m+PHjs2TJktLyyiuvNOr2AACA8rmHEwAAAOtVWVmZj370o0mSPfbYI48//ni++93v5vOf/3xWrFiRxYsX1/uV04IFC1JTU5MkqampyWOPPVZvfQsWLCj1rU9VVVWqqqoaeE8AAIDG5BdOAAAAbLDVq1dn+fLl2WOPPdKyZcvce++9pb45c+Zk3rx5GTJkSJJkyJAhefrpp7Nw4cLSmBkzZqS6ujr9+/d/3+cOAAA0Hr9wAgAAYJ3Gjx+fgw8+OL169crrr7+e6dOn57777stdd92VDh06ZPTo0Rk3blw6deqU6urqnHHGGRkyZEj22muvJMnw4cPTv3//HHfccbnoootSW1ubc889N2PHjvULJgAA+IAROAEAALBOCxcuzPHHH5/58+enQ4cOGTBgQO6666586lOfSpJceumladasWY488sgsX748I0aMyBVXXFF6fvPmzXPbbbfltNNOy5AhQ9K2bduMGjUqEydObKpdAgAAGonACQAAgHW65ppr3rW/VatWmTx5ciZPnrzeMb17984dd9zR0FMDAAA2M+7hBAAAAAAAQFkETgAAAAAAAJRF4AQAAAAAAEBZBE4AAAAAAACUReAEAAAAAABAWQROAAAAAAAAlEXgBAAAAAAAQFkETgAAAAAAAJRF4AQAAAAAAEBZBE4AAAAAAACUReAEAAAAAABAWQROAAAAAAAAlEXgBAAAAAAAQFkETgAAAAAAAJRF4AQAAAAAAEBZBE4AAAAAAACUReAEAAAAAABAWQROAAAAAAAAlEXgBAAAAAAAQFkETgAAAAAAAJRF4AQAAAAAAEBZBE4AAAAAAACUReAEAAAAAABAWQROAAAAAAAAlEXgBAAAAAAAQFk2OnB64IEHcuihh6ZHjx6pqKjIrbfeWq+/KIqcd9556d69e1q3bp1hw4blhRdeqDfm73//e4455phUV1enY8eOGT16dJYuXVpvzFNPPZV99tknrVq1Ss+ePXPRRRdt/N4BAAAAAADQ6DY6cHrjjTcycODATJ48eZ39F110US6//PJcddVVefTRR9O2bduMGDEib731VmnMMccck2effTYzZszIbbfdlgceeCBjxowp9dfV1WX48OHp3bt3Zs2alYsvvjgTJkzI1VdfvQm7CAAAAAAAQGNqsbFPOPjgg3PwwQevs68oilx22WU599xzc9hhhyVJrr/++nTr1i233nprvvCFL+T555/PnXfemccffzx77rlnkuR73/teDjnkkHznO99Jjx49Mm3atKxYsSLXXnttKisrs/POO2f27Nm55JJL6gVTAAAAAAAANL0GvYfT3LlzU1tbm2HDhpXaOnTokMGDB2fmzJlJkpkzZ6Zjx46lsClJhg0blmbNmuXRRx8tjdl3331TWVlZGjNixIjMmTMn//jHP9a57eXLl6eurq7eAgAAAAAAQONr0MCptrY2SdKtW7d67d26dSv11dbWpmvXrvX6W7RokU6dOtUbs651vHMb/2rSpEnp0KFDaenZs2f5OwQAAAAAAMB7atDAqSmNHz8+S5YsKS2vvPJKU08JAAAAAADgQ6FBA6eampokyYIFC+q1L1iwoNRXU1OThQsX1utftWpV/v73v9cbs651vHMb/6qqqirV1dX1FgAAAAAAABpfgwZOffv2TU1NTe69995SW11dXR599NEMGTIkSTJkyJAsXrw4s2bNKo359a9/ndWrV2fw4MGlMQ888EBWrlxZGjNjxozsuOOO2WqrrRpyygAAAAAAAJRpowOnpUuXZvbs2Zk9e3aSZO7cuZk9e3bmzZuXioqKfOlLX8o3vvGN/OIXv8jTTz+d448/Pj169Mjhhx+eJOnXr18OOuignHzyyXnsscfy0EMP5fTTT88XvvCF9OjRI0ly9NFHp7KyMqNHj86zzz6bG2+8Md/97nczbty4BttxAAAAAAAAGkaLjX3CE088kf3337/0eE0INGrUqEydOjVf/epX88Ybb2TMmDFZvHhx9t5779x5551p1apV6TnTpk3L6aefngMPPDDNmjXLkUcemcsvv7zU36FDh9x9990ZO3Zs9thjj3Tp0iXnnXdexowZU86+AgAAAAAA0Ag2OnAaOnRoiqJYb39FRUUmTpyYiRMnrndMp06dMn369HfdzoABA/Lggw9u7PQAAAAAAAB4nzXoPZwAAAAAAAD48BE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAsE6TJk3Kxz72sbRv3z5du3bN4Ycfnjlz5tQb89Zbb2Xs2LHp3Llz2rVrlyOPPDILFiyoN2bevHkZOXJk2rRpk65du+bss8/OqlWr3s9dAQAAGpnACQAAgHW6//77M3bs2DzyyCOZMWNGVq5cmeHDh+eNN94ojTnrrLPyy1/+MjfddFPuv//+vPrqqzniiCNK/W+//XZGjhyZFStW5OGHH851112XqVOn5rzzzmuKXQIAABpJi6aeAAAAAJunO++8s97jqVOnpmvXrpk1a1b23XffLFmyJNdcc02mT5+eAw44IEkyZcqU9OvXL4888kj22muv3H333Xnuuedyzz33pFu3bhk0aFAuuOCCnHPOOZkwYUIqKyubYtcAAIAG5hdOAAAAbJAlS5YkSTp16pQkmTVrVlauXJlhw4aVxuy0007p1atXZs6cmSSZOXNmdt1113Tr1q00ZsSIEamrq8uzzz67zu0sX748dXV19RYAAGDzJnACAADgPa1evTpf+tKX8slPfjK77LJLkqS2tjaVlZXp2LFjvbHdunVLbW1tacw7w6Y1/Wv61mXSpEnp0KFDaenZs2cD7w0AANDQBE4AAAC8p7Fjx+aZZ57JDTfc0OjbGj9+fJYsWVJaXnnllUbfJgAAUB73cAIAAOBdnX766bntttvywAMPZJtttim119TUZMWKFVm8eHG9XzktWLAgNTU1pTGPPfZYvfUtWLCg1LcuVVVVqaqqauC9AAAAGpNfOAEAALBORVHk9NNPzy233JJf//rX6du3b73+PfbYIy1btsy9995bapszZ07mzZuXIUOGJEmGDBmSp59+OgsXLiyNmTFjRqqrq9O/f//3Z0cAAIBG5xdOAAAArNPYsWMzffr0/PznP0/79u1L91zq0KFDWrdunQ4dOmT06NEZN25cOnXqlOrq6pxxxhkZMmRI9tprryTJ8OHD079//xx33HG56KKLUltbm3PPPTdjx471KyYAAPgAafBfOPXp0ycVFRVrLWPHjk2SDB06dK2+U089td465s2bl5EjR6ZNmzbp2rVrzj777KxataqhpwoAAMC7uPLKK7NkyZIMHTo03bt3Ly033nhjacyll16aT3/60znyyCOz7777pqamJjfffHOpv3nz5rntttvSvHnzDBkyJMcee2yOP/74TJw4sSl2CQAAaCQN/gunxx9/PG+//Xbp8TPPPJNPfepT+fd///dS28knn1yvuGjTpk3p32+//XZGjhyZmpqaPPzww5k/f36OP/74tGzZMt/61rcaeroAAACsR1EU7zmmVatWmTx5ciZPnrzeMb17984dd9zRkFMDAAA2Mw0eOG299db1Hl944YXZbrvtst9++5Xa2rRps96bw95999157rnncs8996Rbt24ZNGhQLrjggpxzzjmZMGFCKisrG3rKAAAAAAAAlKHBL6n3TitWrMhPfvKTnHjiiamoqCi1T5s2LV26dMkuu+yS8ePHZ9myZaW+mTNnZtddd023bt1KbSNGjEhdXV2effbZ9W5r+fLlqaurq7cAAAAAAADQ+Br8F07vdOutt2bx4sU54YQTSm1HH310evfunR49euSpp57KOeeckzlz5pSu8V1bW1svbEpSerzmBrXrMmnSpJx//vkNvxMAAAAAAAC8q0YNnK655pocfPDB6dGjR6ltzJgxpX/vuuuu6d69ew488MC89NJL2W677TZ5W+PHj8+4ceNKj+vq6tKzZ89NXh8AAAAAAAAbptECpz//+c+55557Sr9cWp/BgwcnSV588cVst912qampyWOPPVZvzIIFC5Jkvfd9SpKqqqpUVVWVOWsAAAAAAAA2VqPdw2nKlCnp2rVrRo4c+a7jZs+enSTp3r17kmTIkCF5+umns3DhwtKYGTNmpLq6Ov3792+s6QIAAAAAALCJGuUXTqtXr86UKVMyatSotGjx/23ipZdeyvTp03PIIYekc+fOeeqpp3LWWWdl3333zYABA5Ikw4cPT//+/XPcccfloosuSm1tbc4999yMHTvWL5gAAAAAAAA2Q40SON1zzz2ZN29eTjzxxHrtlZWVueeee3LZZZfljTfeSM+ePXPkkUfm3HPPLY1p3rx5brvttpx22mkZMmRI2rZtm1GjRmXixImNMVUAAAAAAADK1CiB0/Dhw1MUxVrtPXv2zP333/+ez+/du3fuuOOOxpgaAAAAAAAADazR7uEEAAAAAADAh4PACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoS4umngAAAAAAQEOZN29eFi1a1NTToAl06dIlvXr1auppwIeWwAkAoiD5MFOQAADAB8e8efPSb6edsuzNN5t6KjSBNq1b5/k//EGNB01E4ATAh968efOyU79+eXPZsqaeCk2gdZs2+cPzzytIAADgA2DRokVZ9uabufqII7JDly5NPR3eR39ctChjbr45ixYtUt9BExE4AfCht2jRory5bFk+940r07Xv9k09Hd5HC+e+kJ+ee5qCBAAAPmB26NIlg3r0aOppAHyoCJwA4P+va9/t85F+A5t6GgAAAACwxWnW1BMAAAAAAABgyyZwAgAAAAAAoCwCJwAAAAAAAMoicAIAAAAAAKAsAicAAAAAAADK0uCB04QJE1JRUVFv2WmnnUr9b731VsaOHZvOnTunXbt2OfLII7NgwYJ665g3b15GjhyZNm3apGvXrjn77LOzatWqhp4qAAAAAAAADaBFY6x05513zj333PP/baTF/7eZs846K7fffntuuummdOjQIaeffnqOOOKIPPTQQ0mSt99+OyNHjkxNTU0efvjhzJ8/P8cff3xatmyZb33rW40xXQAAAAAAAMrQKIFTixYtUlNTs1b7kiVLcs0112T69Ok54IADkiRTpkxJv3798sgjj2SvvfbK3Xffneeeey733HNPunXrlkGDBuWCCy7IOeeckwkTJqSysrIxpgwAAAAAAMAmapR7OL3wwgvp0aNHtt122xxzzDGZN29ekmTWrFlZuXJlhg0bVhq70047pVevXpk5c2aSZObMmdl1113TrVu30pgRI0akrq4uzz77bGNMFwAAAAAAgDI0+C+cBg8enKlTp2bHHXfM/Pnzc/7552efffbJM888k9ra2lRWVqZjx471ntOtW7fU1tYmSWpra+uFTWv61/Stz/Lly7N8+fLS47q6ugbaIwAAAAAAAN5NgwdOBx98cOnfAwYMyODBg9O7d+/89Kc/TevWrRt6cyWTJk3K+eef32jrBwAAAAAAYN0a5ZJ679SxY8fssMMOefHFF1NTU5MVK1Zk8eLF9cYsWLCgdM+nmpqaLFiwYK3+NX3rM378+CxZsqS0vPLKKw27IwAAAAAAAKxTowdOS5cuzUsvvZTu3btnjz32SMuWLXPvvfeW+ufMmZN58+ZlyJAhSZIhQ4bk6aefzsKFC0tjZsyYkerq6vTv33+926mqqkp1dXW9BQAAAAAAgMbX4JfU+8pXvpJDDz00vXv3zquvvpqvf/3rad68eY466qh06NAho0ePzrhx49KpU6dUV1fnjDPOyJAhQ7LXXnslSYYPH57+/fvnuOOOy0UXXZTa2tqce+65GTt2bKqqqhp6ugAAAAAAAJSpwQOnv/zlLznqqKPy2muvZeutt87ee++dRx55JFtvvXWS5NJLL02zZs1y5JFHZvny5RkxYkSuuOKK0vObN2+e2267LaeddlqGDBmStm3bZtSoUZk4cWJDTxUAAAAAAIAG0OCB0w033PCu/a1atcrkyZMzefLk9Y7p3bt37rjjjoaeGgAAAAAAAI2g0e/hBAAAAAAAwAebwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAIB1euCBB3LooYemR48eqaioyK233lqvvyiKnHfeeenevXtat26dYcOG5YUXXqg35u9//3uOOeaYVFdXp2PHjhk9enSWLl36Pu4FAADwfhA4AQAAsE5vvPFGBg4cmMmTJ6+z/6KLLsrll1+eq666Ko8++mjatm2bESNG5K233iqNOeaYY/Lss89mxowZue222/LAAw9kzJgx79cuAAAA75MWTT0BAAAANk8HH3xwDj744HX2FUWRyy67LOeee24OO+ywJMn111+fbt265dZbb80XvvCFPP/887nzzjvz+OOPZ88990ySfO9738shhxyS73znO+nRo8f7ti8AAEDj8gsnAAAANtrcuXNTW1ubYcOGldo6dOiQwYMHZ+bMmUmSmTNnpmPHjqWwKUmGDRuWZs2a5dFHH33f5wwAADQev3ACAABgo9XW1iZJunXrVq+9W7dupb7a2tp07dq1Xn+LFi3SqVOn0ph1Wb58eZYvX156XFdX11DTBgAAGolfOAEAALBZmTRpUjp06FBaevbs2dRTAgAA3oPACQAAgI1WU1OTJFmwYEG99gULFpT6ampqsnDhwnr9q1atyt///vfSmHUZP358lixZUlpeeeWVBp49AADQ0AROAAAAbLS+ffumpqYm9957b6mtrq4ujz76aIYMGZIkGTJkSBYvXpxZs2aVxvz617/O6tWrM3jw4PWuu6qqKtXV1fUWAABg8+YeTgAAAKzT0qVL8+KLL5Yez507N7Nnz06nTp3Sq1evfOlLX8o3vvGNbL/99unbt2/+53/+Jz169Mjhhx+eJOnXr18OOuignHzyybnqqquycuXKnH766fnCF76QHj16NNFeAQAAjUHgBAAAwDo98cQT2X///UuPx40blyQZNWpUpk6dmq9+9at54403MmbMmCxevDh777137rzzzrRq1ar0nGnTpuX000/PgQcemGbNmuXII4/M5Zdf/r7vCwAA0LgETgAAAKzT0KFDUxTFevsrKioyceLETJw4cb1jOnXqlOnTpzfG9AAAgM2IezgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGUROAEAAAAAAFAWgRMAAAAAAABlETgBAAAAAABQFoETAAAAAAAAZRE4AQAAAAAAUBaBEwAAAAAAAGVp8MBp0qRJ+djHPpb27duna9euOfzwwzNnzpx6Y4YOHZqKiop6y6mnnlpvzLx58zJy5Mi0adMmXbt2zdlnn51Vq1Y19HQBAAAAAAAoU4uGXuH999+fsWPH5mMf+1hWrVqVr33taxk+fHiee+65tG3btjTu5JNPzsSJE0uP27RpU/r322+/nZEjR6ampiYPP/xw5s+fn+OPPz4tW7bMt771rYaeMgAAAAAAAGVo8MDpzjvvrPd46tSp6dq1a2bNmpV999231N6mTZvU1NSscx133313nnvuudxzzz3p1q1bBg0alAsuuCDnnHNOJkyYkMrKyoaeNgAAAAAAAJuo0e/htGTJkiRJp06d6rVPmzYtXbp0yS677JLx48dn2bJlpb6ZM2dm1113Tbdu3UptI0aMSF1dXZ599tnGnjIAAAAAAAAbocF/4fROq1evzpe+9KV88pOfzC677FJqP/roo9O7d+/06NEjTz31VM4555zMmTMnN998c5Kktra2XtiUpPS4trZ2ndtavnx5li9fXnpcV1fX0LsDAAAAAADAOjRq4DR27Ng888wz+e1vf1uvfcyYMaV/77rrrunevXsOPPDAvPTSS9luu+02aVuTJk3K+eefX9Z8AQAAAAAA2HiNdkm9008/Pbfddlt+85vfZJtttnnXsYMHD06SvPjii0mSmpqaLFiwoN6YNY/Xd9+n8ePHZ8mSJaXllVdeKXcXAAAAAAAA2AANHjgVRZHTTz89t9xyS37961+nb9++7/mc2bNnJ0m6d++eJBkyZEiefvrpLFy4sDRmxowZqa6uTv/+/de5jqqqqlRXV9dbAAAAAAAAaHwNfkm9sWPHZvr06fn5z3+e9u3bl+651KFDh7Ru3TovvfRSpk+fnkMOOSSdO3fOU089lbPOOiv77rtvBgwYkCQZPnx4+vfvn+OOOy4XXXRRamtrc+6552bs2LGpqqpq6CkDAAAAAABQhgb/hdOVV16ZJUuWZOjQoenevXtpufHGG5MklZWVueeeezJ8+PDstNNO+fKXv5wjjzwyv/zlL0vraN68eW677bY0b948Q4YMybHHHpvjjz8+EydObOjpAgAAAAAAUKYG/4VTURTv2t+zZ8/cf//977me3r1754477mioaQEAAAAAANBIGvwXTgAAAAAAAHy4CJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAACiLwAkAAAAAAICyCJwAAAAAAAAoi8AJAAAAAACAsgicAAAAAAAAKIvACQAAAAAAgLIInAAAAAAAAChLi6aeAAAAjWPevHlZtGhRU0+DJtClS5f06tWrqacBAAA0ALXdh9uWVN8JnAAAPoDmzZuXnfr1y5vLljX1VGgCrdu0yR+ef36LKUoAAIB1mzdvXvrttFOWvflmU0+FJtKmdes8/4c/bBH1ncAJAOADaNGiRXlz2bJ87htXpmvf7Zt6OryPFs59IT8997QsWrRoiyhIAACA9Vu0aFGWvflmrj7iiOzQpUtTT4f32R8XLcqYm2/eYuo7gRMAwAdY177b5yP9Bjb1NAAAACjDDl26ZFCPHk09DXhXzZp6Au9m8uTJ6dOnT1q1apXBgwfnsccea+opAQAAsAnUdwAA8MG22QZON954Y8aNG5evf/3r+d3vfpeBAwdmxIgRWbhwYVNPDQAAgI2gvgMAgA++zTZwuuSSS3LyySfnP/7jP9K/f/9cddVVadOmTa699tqmnhoAAAAbQX0HAAAffJvlPZxWrFiRWbNmZfz48aW2Zs2aZdiwYZk5c+Y6n7N8+fIsX7689HjJkiVJkrq6usad7LtYunRpkuSvzz+VFcveaLJ50DT+9ueXkvzzOGjK47ChOa4/vD6ox3TiuP4wc1zzQbS5HNdrtl0URZPNgc3DB62+e3L+/LyxYkWTzYP334uvvZak6c+rjcFx/eHluOaD6IN6XDumP9w2h+N6Y2q7imIzrABfffXVfOQjH8nDDz+cIUOGlNq/+tWv5v7778+jjz661nMmTJiQ888///2cJgAA8B5eeeWVbLPNNk09DZqQ+g4AALZ8G1LbbZa/cNoU48ePz7hx40qPV69enb///e/p3LlzKioqmnBmH151dXXp2bNnXnnllVRXVzf1dKBBOK75IHJc80HjmN48FEWR119/PT169GjqqbAFUt9tfpxb+SByXPNB45jmg8hx3fQ2prbbLAOnLl26pHnz5lmwYEG99gULFqSmpmadz6mqqkpVVVW9to4dOzbWFNkI1dXVTgZ84Diu+SByXPNB45hueh06dGjqKbAZUN99sDi38kHkuOaDxjHNB5HjumltaG3XrJHnsUkqKyuzxx575N577y21rV69Ovfee2+9SzAAAACweVPfAQDAh8Nm+QunJBk3blxGjRqVPffcMx//+Mdz2WWX5Y033sh//Md/NPXUAAAA2AjqOwAA+ODbbAOnz3/+8/nb3/6W8847L7W1tRk0aFDuvPPOdOvWramnxgaqqqrK17/+9bUuhQFbMsc1H0SOaz5oHNOw+VHfbfmcW/kgclzzQeOY5oPIcb1lqSiKomjqSQAAAAAAALDl2izv4QQAAAAAAMCWQ+AEAAAAAABAWQROAAAAAAAAlEXgBOvw8ssvp6KiIrNnz27qqbCJ+vTpk8suu6z0uKKiIrfeeut6x3vN39t9992XioqKLF68uKmnssUaOnRovvSlL5Ue/+txui7vdex+2Hnvvn+cA7ZMEyZMyKBBg5p6GgBNyueFLZ/6ruH5bFc+9V3D8959/zgHbJm2hPpO4LQZO+GEE1JRUZELL7ywXvutt96aioqKBt1WRUVFaWnbtm223377nHDCCZk1a1aDbqcx+Y9S4zv00ENz0EEHrbPvwQcfTEVFRZ566qlG2fbQoUPX+X5IkpEjR6aioiITJkwotT3++OMZM2bMJm9vzfG0Zmnfvn123nnnjB07Ni+88MImr/f95gNE41tzrj711FPX6hs7dmwqKipywgknJEluvvnmXHDBBWVtz/ma9dmYY3Fzt6kfoqdOnVp6fzRv3jxbbbVVBg8enIkTJ2bJkiUNP9FGsiUUEcDGU99tHJ8XGp/6Tn3H2tR3G8/5unGo79R3WyqB02auVatW+fa3v51//OMfjb6tKVOmZP78+Xn22WczefLkLF26NIMHD87111/f6NtmyzB69OjMmDEjf/nLX9bqmzJlSvbcc88MGDBgo9a5YsWKDR7bs2fPTJ06tV7bX//619x7773p3r17vfatt946bdq02ai5rMs999yT+fPn58knn8y3vvWtPP/88xk4cGDuvffestfNB0fPnj1zww035M033yy1vfXWW5k+fXp69epVauvUqVPat29f9vacr1mfDT0Wm8rGnPM3VXV1debPn5+//OUvefjhhzNmzJhcf/31GTRoUF599dVG3z7Au1HfsTlR36nvWDf1HZsL9Z36bkskcNrMDRs2LDU1NZk0adJ6x/zf//1fdt5551RVVaVPnz753//933r9ffr0ybe+9a2ceOKJad++fXr16pWrr756rfV07NgxNTU16dOnT4YPH56f/exnOeaYY3L66afXK4h++9vfZp999knr1q3Ts2fPnHnmmXnjjTfqbe8b3/hGjj/++LRr1y69e/fOL37xi/ztb3/LYYcdlnbt2mXAgAF54okn6m1/Q9b7bvvRt2/fJMluu+2WioqKDB06tNT3ox/9KP369UurVq2y00475Yorrqi37cceeyy77bZbWrVqlT333DO///3v1/v3/jD79Kc/na233nqtomDp0qW56aabMnr06A16HS+44IIcf/zxqa6uzpgxY3LAAQfk9NNPr7fOv/3tb6msrKz3wf/Tn/50Fi1alIceeqjUdt1112X48OHp2rVrvee/10/ZN/Q179y5c2pqarLtttvmsMMOyz333JPBgwdn9OjRefvtt0vjfv7zn2f33XdPq1atsu222+b888/PqlWrSv0VFRX5wQ9+kE9/+tNp06ZN+vXrl5kzZ+bFF1/M0KFD07Zt23ziE5/ISy+9VG/7G7LeH/3oR/nMZz6TNm3aZPvtt88vfvGLJP/8ltH++++fJNlqq63qfftl9erVmTRpUvr27ZvWrVtn4MCB+dnPflZv23fccUd22GGHtG7dOvvvv39efvnl9f49P+x233339OzZMzfffHOp7eabb06vXr2y2267ldr+9ZIL/+qFF17Ivvvum1atWqV///6ZMWPGOsc5Xztfr8+GHovLly/PmWeema5du6ZVq1bZe++98/jjj9db14acAzblnJ8k55xzTnbYYYe0adMm2267bf7nf/4nK1euTPLPb7Gdf/75efLJJ0vfZlvz353FixfnpJNOytZbb53q6uoccMABefLJJ+vNqaKiIjU1NenevXv69euX0aNH5+GHH87SpUvz1a9+tTTuvc6Da75BfNddd2W33XZL69atc8ABB2ThwoX51a9+lX79+qW6ujpHH310li1bttHrvffee7PnnnumTZs2+cQnPpE5c+Y0yP5feOGF6datW9q3b5/Ro0fnrbfeWut1A5qO+s7nhc2J+k59x7qp75yvNxfqO/XdFlnfFWy2Ro0aVRx22GHFzTffXLRq1ap45ZVXiqIoiltuuaVY89I98cQTRbNmzYqJEycWc+bMKaZMmVK0bt26mDJlSmk9vXv3Ljp16lRMnjy5eOGFF4pJkyYVzZo1K/7whz+UxiQpbrnllrXm8Pvf/75IUtx4441FURTFiy++WLRt27a49NJLiz/+8Y/FQw89VOy2227FCSecsNb2rrrqquKPf/xjcdpppxXV1dXFQQcdVPz0pz8t5syZUxx++OFFv379itWrV2/0ete3H4899liRpLjnnnuK+fPnF6+99lpRFEXxk5/8pOjevXvxf//3f8Wf/vSn4v/+7/+KTp06FVOnTi2Koihef/31Yuutty6OPvro4plnnil++ctfFttuu22RpPj9739f5qv4wXP22WcX2223Xem1K4qiuPbaa4vWrVsXs2fP3qDXsbq6uvjOd75TvPjii8WLL75YTJs2rdhqq62Kt956qzTukksuKfr06VPazn777Vd88YtfLM4888xi9OjRpXHbb799ccsttxQDBw4svv71r9fbzqWXXlp6/M5jfENe87lz5673GFjzHnz00UeLoiiKBx54oKiuri6mTp1avPTSS8Xdd99d9OnTp5gwYUK97X/kIx8pbrzxxtJ7oE+fPsUBBxxQ3HnnncVzzz1X7LXXXsVBBx1Ues6GrnebbbYppk+fXrzwwgvFmWeeWbRr16547bXXilWrVhX/93//VyQp5syZU8yfP79YvHhxURRF8Y1vfKPYaaedijvvvLN46aWXiilTphRVVVXFfffdVxRFUcybN6+oqqoqxo0bV/zhD38ofvKTnxTdunUrkhT/+Mc/3u0Q+dBZc66+5JJLigMPPLDUfuCBBxaXXnppcdhhhxWjRo0qiuL/O47XeOdx+vbbbxe77LJLceCBBxazZ88u7r///mK33XZb6/zsfO18vT4bcyyeeeaZRY8ePYo77rijePbZZ4tRo0YVW221Vem12JBzwIYeC/96zi+KorjggguKhx56qJg7d27xi1/8oujWrVvx7W9/uyiKoli2bFnx5S9/udh5552L+fPnF/Pnzy+WLVtWFEVRDBs2rDj00EOLxx9/vPjjH/9YfPnLXy46d+5cmveUKVOKDh06rPPv88UvfrFo3759sWrVqqIo3vs8+Jvf/KZIUuy1117Fb3/72+J3v/td8dGPfrTYb7/9iuHDhxe/+93vigceeKDo3LlzceGFF5a2s6HrHTx4cHHfffcVzz77bLHPPvsUn/jEJ8re/xtvvLGoqqoqfvSjHxV/+MMfiv/+7/8u2rdvXwwcOHAjjyagMajvfF7YHKnv1Hfqu/rUd87Xmwv1nfpuS63vBE6bsTUnlqIoir322qs48cQTi6KoX5AcffTRxac+9al6zzv77LOL/v37lx737t27OPbYY0uPV69eXXTt2rW48sorS23r+w/cm2++WSQpnSRGjx5djBkzpt6YBx98sGjWrFnx5ptvrnN78+fPL5IU//M//1NqmzlzZpGkmD9//iav91/3Y30fILfbbrti+vTp9douuOCCYsiQIUVRFMUPfvCDonPnzqXtFEVRXHnllf4Dtx7PP/98kaT4zW9+U2rbZ599imOPPXaDX8fDDz+83pg333yz2GqrrUofpIqiKAYMGFDvg/eaD3KzZ88u2rdvXyxdurS4//77i65duxYrV67cqIJkQ17zdytI1vwN1sz3wAMPLL71rW/VG/PjH/+46N69e73tn3vuuaXHa94D11xzTant//2//1e0atWq9HhT1rt06dIiSfGrX/2qKIr/7z987ywi3nrrraJNmzbFww8/XG/do0ePLo466qiiKIpi/Pjx9c4jRVEU55xzjoJkHdacqxcuXFhUVVUVL7/8cvHyyy8XrVq1Kv72t79tcEFy1113FS1atCj++te/lvp/9atfbXBB4nxdf10fRht6LC5durRo2bJlMW3atNJzV6xYUfTo0aO46KKLiqLYsHPApp7z1+Xiiy8u9thjj9Ljr3/962t9iH7wwQeL6urqev/zqij+edz84Ac/KIri3QuSNcfKggULNug8uOb8ec8995T6J02aVCQpXnrppVLbKaecUowYMaIoig07v65rvbfffnuRpPR329T9HzJkSPGf//mf9foHDx682Rck8GGhvvN5YXOkvlPfqe/qU985X28u1Hfquy21vmux/t8+sTn59re/nQMOOCBf+cpX6rU///zzOeyww+q1ffKTn8xll12Wt99+O82bN0+SetddXvNTxIULF77ndouiKD0nSZ588sk89dRTmTZtWr0xq1evzty5c9OvX7+1ttetW7ckya677rpW28KFC1NTU7NJ692Q/XjjjTfy0ksvZfTo0Tn55JNL7atWrUqHDh2S/PNvOGDAgLRq1arUP2TIkPf823xY7bTTTvnEJz6Ra6+9NkOHDs2LL76YBx98MBMnTszZZ5+9Qa/jnnvuWW+drVq1ynHHHZdrr702n/vc5/K73/0uzzzzTOnSAe80cODAbL/99vnZz36W3/zmNznuuOPSosXGncrKfc3X9b546KGH8s1vfrM05u23385bb72VZcuWla41viHvi7feeit1dXWprq7epPW2bds21dXV7/q+ePHFF7Ns2bJ86lOfqte+YsWK0k+yn3/++QwePLhev/fFu9t6660zcuTITJ06NUVRZOTIkenSpcsGP//5559Pz54906NHj1Jbucel8/WH03sdiy+99FJWrlyZT37yk6W2li1b5uMf/3ief/75JBt2DtjQY+Ffz/lJcuONN+byyy/PSy+9lKVLl2bVqlWprq5+1/168skns3Tp0nTu3Lle+5tvvrnW5WrW5Z3vkQ05D67xr++RNZeJeGfbY489lmTDzq/rWu+a+1QsXLhwvddi35D9f/7559e6qfCQIUPym9/8Zp3rBJqO+s7nhc2F+k59x7qp75yvNxfqu3VT322+BE5biH333TcjRozI+PHjS9fo3RgtW7as97iioiKrV69+z+etOTGtuR7r0qVLc8opp+TMM89ca+w730Dv3N6a/ziuq23NHDZlvRuyH0uXLk2S/PCHP1zrxLqmWGPjjR49OmeccUYmT56cKVOmZLvttst+++23wa9j27Zt1+o/6aSTMmjQoPzlL3/JlClTcsABB6R3797r3P6JJ56YyZMn57nnniv9R+D9tK73xfnnn58jjjhirbHv/OC0Ke+LjV3vmvVsyPvi9ttvz0c+8pF6fVVVVet9Hu/txBNPLF2vfvLkye/rtp2veaf341jc1HP+zJkzc8wxx+T888/PiBEj0qFDh9xwww1r3aNkXdvr3r177rvvvrX6Onbs+J7zff7551NdXZ3OnTvnT3/6U5INOw/+6/vh3Y7tjTm/vtv7bF3K3X9g86K+27T98Hmhcajv1Hesm/rO+Xpzob5bm/pu8yVw2oJceOGFGTRoUHbcccdSW79+/erdYDNJHnrooeywww4NcgK/7LLLUl1dnWHDhiX5583qnnvuuXz0ox8te93v1BDrraysTJJ6N/rs1q1bevTokT/96U855phj1vm8fv365cc//nHeeuut0oe8Rx55ZJPn8WHwuc99Ll/84hczffr0XH/99TnttNNSUVFR1uu46667Zs8998wPf/jDTJ8+Pd///vfXO/boo4/OV77ylQwcODD9+/ff6G2V85qvXr06l19+efr27Vv6NsPuu++eOXPmNMr7otz1rut90b9//1RVVWXevHnZb7/91vm8fv36rfUNRO+L93bQQQdlxYoVqaioyIgRIzbquf369csrr7yS+fPnl74NszF/c+dr3undjsXtttsulZWVeeihh0r/42flypV5/PHHSzc93pBzwKYeCw8//HB69+6d//7v/y61/fnPf643prKyst7xsWZ7tbW1adGiRfr06bNR21y4cGGmT5+eww8/PM2aNdug8+CmaKj1bur+9+vXL48++miOP/74Upv3CGy+1HfvzueF94/6Tn3HuqnvNozzdeNT39Wnvtu8CZy2ILvuumuOOeaYXH755aW2L3/5y/nYxz6WCy64IJ///Oczc+bMfP/7388VV1yx0etfvHhxamtrs3z58vzxj3/MD37wg9x66625/vrrS8nqOeeck7322iunn356TjrppLRt2zbPPfdcZsyY8a4fIN9LQ6y3a9euad26de68885ss802adWqVTp06JDzzz8/Z555Zjp06JCDDjooy5cvzxNPPJF//OMfGTduXI4++uj893//d04++eSMHz8+L7/8cr7zne9s8r58GLRr1y6f//znM378+NTV1ZW+lVnu63jSSSfl9NNPT9u2bfOZz3xmveO22mqrzJ8/f61vIWyojXnNX3vttdTW1mbZsmV55plnctlll+Wxxx7L7bffXir6zzvvvHz6059Or1698tnPfjbNmjXLk08+mWeeeSbf+MY3NmmODbXe3r17p6KiIrfddlsOOeSQtG7dOu3bt89XvvKVnHXWWVm9enX23nvvLFmyJA899FCqq6szatSonHrqqfnf//3fnH322TnppJMya9asTJ06dZP35cOiefPmpW+ibez/FBo2bFh22GGHjBo1KhdffHHq6urqfWB7J+dr5+v38m7HYtu2bXPaaafl7LPPTqdOndKrV69cdNFFWbZsWUaPHp0kG3QO2NRjYfvtt8+8efNyww035GMf+1huv/323HLLLfXG9OnTJ3Pnzs3s2bOzzTbbpH379hk2bFiGDBmSww8/PBdddFF22GGHvPrqq7n99tvzmc98pnRph6IoUltbm6Iosnjx4sycOTPf+ta30qFDh1x44YVJskHnwU3RUOvd1P3/4he/mBNOOCF77rlnPvnJT2batGl59tln610iAth8qO/enc8L7x/1nfqOdVPfOV9vLtR36rstqr57f24VxaZ4501l15g7d25RWVlZvPOl+9nPflb079+/aNmyZdGrV6/i4osvrvecf725ZlEUa92AM0lpadWqVbHddtsVo0aNKmbNmrXWvB577LHiU5/6VNGuXbuibdu2xYABA4pvfvOb77q9/MtNENd1Q8FNWe+/7scPf/jDomfPnkWzZs2K/fbbr9Q+bdq0YtCgQUVlZWWx1VZbFfvuu29x8803l/pnzpxZDBw4sKisrCwGDRpU/N///Z+bFL6Hhx9+uEhSHHLIIfXaN+V1XOP1118v2rRps9YN8Ypi7Ztx/quNualsUbz3a77mGF2ztGnTpujXr1/xn//5n8ULL7yw1vbvvPPO4hOf+ETRunXrorq6uvj4xz9eXH311evd/rreA+u6AezGrrcoiqJDhw7FlClTSo8nTpxY1NTUFBUVFaWbm65evbq47LLLih133LFo2bJlsfXWWxcjRowo7r///tLzfvnLXxYf/ehHi6qqqmKfffYprr32WjeVXYd1navfaUNvKlsURTFnzpxi7733LiorK4sddtihuPPOO9d5U1nna+frddmYY/HNN98szjjjjKJLly5FVVVV8clPfrJ47LHH6o3fkHPApp7zzz777KJz585Fu3btis9//vPFpZdeWu9msG+99VZx5JFHFh07diySlM5pdXV1xRlnnFH06NGjaNmyZdGzZ8/imGOOKebNm1cUxT9vKrvm/VFRUVF06NCh+PjHP15MnDixWLJkSb05vNd5cF3n5HXdtPZfbwC7Kev9/e9/XyQp5s6dW9b+F0VRfPOb3yy6dOlStGvXrhg1alTx1a9+dbO/qSx8WKjvfF7YnKnv6lPffXip75yvNxfqO/VdUWyZ9V1FUfz/77AF0MRefvnlbLfddnn88cez++67N/V0AAAA2ETqOwD48BE4AU1u5cqVee211/KVr3wlc+fOXeu69QAAAGwZ1HcA8OHVrKknAPDQQw+le/fuefzxx3PVVVc19XQAAADYROo7APjw8gsnAAAAAAAAyuIXTgAAAAAAAJRF4AQAAAAAAEBZBE4AAAAAAACUReAEAAAAAABAWQROAAAAAAAAlEXgBAAAAAAAQFkETgAAAAAAAJRF4AQAAAAAAEBZBE4AAAAAAACU5f8Hr6zAJKJAuQgAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Training"
      ],
      "metadata": {
        "id": "3VjJR77OD6KJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "checkpoint_filepath = '/content/drive/MyDrive/blockchain/best_model_alz_test'\n",
        "model_checkpoint = ModelCheckpoint(\n",
        "    checkpoint_filepath,\n",
        "    save_best_only=True,\n",
        "    monitor='val_f1_score',  # Monitor the validation macro-F1 score\n",
        "    mode='max',\n",
        "    verbose=0,\n",
        "    save_weights_only=False  # Save the entire model\n",
        ")\n",
        "\n",
        "# Create and compile the model\n",
        "model = FedAvg(num_classes=4)\n",
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=[tfa.metrics.F1Score(num_classes=4,average='macro')])\n",
        "# Train the model\n",
        "epochs = 80\n",
        "history = model.fit(train_dataset, epochs=epochs, validation_data=test_dataset, callbacks=[model_checkpoint])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bjI-0TAdD5Xs",
        "outputId": "f5e3361f-b694-4da8-d2c8-85ec6b382fb1"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/80\n",
            "128/128 [==============================] - 9s 53ms/step - loss: 1.0307 - f1_score: 0.2537 - val_loss: 0.9472 - val_f1_score: 0.3162\n",
            "Epoch 2/80\n",
            "128/128 [==============================] - 5s 39ms/step - loss: 0.8570 - f1_score: 0.3878 - val_loss: 0.7492 - val_f1_score: 0.4079\n",
            "Epoch 3/80\n",
            "128/128 [==============================] - 5s 36ms/step - loss: 0.6405 - f1_score: 0.5251 - val_loss: 0.5287 - val_f1_score: 0.5869\n",
            "Epoch 4/80\n",
            "128/128 [==============================] - 6s 49ms/step - loss: 0.4724 - f1_score: 0.6352 - val_loss: 0.3416 - val_f1_score: 0.7125\n",
            "Epoch 5/80\n",
            "128/128 [==============================] - 5s 37ms/step - loss: 0.2860 - f1_score: 0.8231 - val_loss: 0.2240 - val_f1_score: 0.9137\n",
            "Epoch 6/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.2250 - f1_score: 0.8842 - val_loss: 0.2266 - val_f1_score: 0.8555\n",
            "Epoch 7/80\n",
            "128/128 [==============================] - 6s 47ms/step - loss: 0.2081 - f1_score: 0.8714 - val_loss: 0.1288 - val_f1_score: 0.9546\n",
            "Epoch 8/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.1848 - f1_score: 0.8924 - val_loss: 0.2400 - val_f1_score: 0.9378\n",
            "Epoch 9/80\n",
            "128/128 [==============================] - 5s 38ms/step - loss: 0.1371 - f1_score: 0.9259 - val_loss: 0.1324 - val_f1_score: 0.9611\n",
            "Epoch 10/80\n",
            "128/128 [==============================] - 5s 36ms/step - loss: 0.1450 - f1_score: 0.9193 - val_loss: 0.0968 - val_f1_score: 0.9768\n",
            "Epoch 11/80\n",
            "128/128 [==============================] - 4s 34ms/step - loss: 0.1111 - f1_score: 0.9447 - val_loss: 0.2286 - val_f1_score: 0.8546\n",
            "Epoch 12/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.1275 - f1_score: 0.9316 - val_loss: 0.1345 - val_f1_score: 0.9215\n",
            "Epoch 13/80\n",
            "128/128 [==============================] - 5s 38ms/step - loss: 0.1207 - f1_score: 0.9537 - val_loss: 0.0877 - val_f1_score: 0.9803\n",
            "Epoch 14/80\n",
            "128/128 [==============================] - 6s 46ms/step - loss: 0.0894 - f1_score: 0.9561 - val_loss: 0.0566 - val_f1_score: 0.9856\n",
            "Epoch 15/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0471 - f1_score: 0.9775 - val_loss: 0.0940 - val_f1_score: 0.9827\n",
            "Epoch 16/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0846 - f1_score: 0.9605 - val_loss: 0.0816 - val_f1_score: 0.9828\n",
            "Epoch 17/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0761 - f1_score: 0.9668 - val_loss: 0.0684 - val_f1_score: 0.9816\n",
            "Epoch 18/80\n",
            "128/128 [==============================] - 3s 26ms/step - loss: 0.0785 - f1_score: 0.9735 - val_loss: 0.0786 - val_f1_score: 0.9781\n",
            "Epoch 19/80\n",
            "128/128 [==============================] - 5s 37ms/step - loss: 0.0561 - f1_score: 0.9726 - val_loss: 0.0773 - val_f1_score: 0.9818\n",
            "Epoch 20/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0535 - f1_score: 0.9781 - val_loss: 0.0861 - val_f1_score: 0.9728\n",
            "Epoch 21/80\n",
            "128/128 [==============================] - 3s 27ms/step - loss: 0.0471 - f1_score: 0.9667 - val_loss: 0.0635 - val_f1_score: 0.9792\n",
            "Epoch 22/80\n",
            "128/128 [==============================] - 6s 50ms/step - loss: 0.0497 - f1_score: 0.9753 - val_loss: 0.0519 - val_f1_score: 0.9870\n",
            "Epoch 23/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0407 - f1_score: 0.9781 - val_loss: 0.0483 - val_f1_score: 0.9869\n",
            "Epoch 24/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0447 - f1_score: 0.9624 - val_loss: 0.0683 - val_f1_score: 0.9834\n",
            "Epoch 25/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0412 - f1_score: 0.9776 - val_loss: 0.0755 - val_f1_score: 0.9818\n",
            "Epoch 26/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0647 - f1_score: 0.9731 - val_loss: 0.0761 - val_f1_score: 0.9810\n",
            "Epoch 27/80\n",
            "128/128 [==============================] - 4s 31ms/step - loss: 0.0643 - f1_score: 0.9663 - val_loss: 0.0653 - val_f1_score: 0.9793\n",
            "Epoch 28/80\n",
            "128/128 [==============================] - 4s 26ms/step - loss: 0.0575 - f1_score: 0.9668 - val_loss: 0.1176 - val_f1_score: 0.9578\n",
            "Epoch 29/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0671 - f1_score: 0.9690 - val_loss: 0.0651 - val_f1_score: 0.9781\n",
            "Epoch 30/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0487 - f1_score: 0.9769 - val_loss: 0.0628 - val_f1_score: 0.9822\n",
            "Epoch 31/80\n",
            "128/128 [==============================] - 5s 35ms/step - loss: 0.0417 - f1_score: 0.9718 - val_loss: 0.0668 - val_f1_score: 0.9865\n",
            "Epoch 32/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0334 - f1_score: 0.9808 - val_loss: 0.0523 - val_f1_score: 0.9853\n",
            "Epoch 33/80\n",
            "128/128 [==============================] - 5s 37ms/step - loss: 0.0206 - f1_score: 0.9912 - val_loss: 0.0353 - val_f1_score: 0.9911\n",
            "Epoch 34/80\n",
            "128/128 [==============================] - 5s 37ms/step - loss: 0.0470 - f1_score: 0.9761 - val_loss: 0.1177 - val_f1_score: 0.9482\n",
            "Epoch 35/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0349 - f1_score: 0.9767 - val_loss: 0.0492 - val_f1_score: 0.9868\n",
            "Epoch 36/80\n",
            "128/128 [==============================] - 3s 26ms/step - loss: 0.0514 - f1_score: 0.9728 - val_loss: 0.0670 - val_f1_score: 0.9793\n",
            "Epoch 37/80\n",
            "128/128 [==============================] - 5s 38ms/step - loss: 0.0263 - f1_score: 0.9887 - val_loss: 0.0704 - val_f1_score: 0.9869\n",
            "Epoch 38/80\n",
            "128/128 [==============================] - 5s 40ms/step - loss: 0.0251 - f1_score: 0.9850 - val_loss: 0.0697 - val_f1_score: 0.9869\n",
            "Epoch 39/80\n",
            "128/128 [==============================] - 5s 37ms/step - loss: 0.0326 - f1_score: 0.9904 - val_loss: 0.0964 - val_f1_score: 0.9713\n",
            "Epoch 40/80\n",
            "128/128 [==============================] - 3s 26ms/step - loss: 0.0538 - f1_score: 0.9730 - val_loss: 0.0687 - val_f1_score: 0.9808\n",
            "Epoch 41/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0315 - f1_score: 0.9829 - val_loss: 0.0641 - val_f1_score: 0.9857\n",
            "Epoch 42/80\n",
            "128/128 [==============================] - 5s 35ms/step - loss: 0.0354 - f1_score: 0.9842 - val_loss: 0.0549 - val_f1_score: 0.9870\n",
            "Epoch 43/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0431 - f1_score: 0.9767 - val_loss: 0.0769 - val_f1_score: 0.9765\n",
            "Epoch 44/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0472 - f1_score: 0.9707 - val_loss: 0.0684 - val_f1_score: 0.9840\n",
            "Epoch 45/80\n",
            "128/128 [==============================] - 4s 31ms/step - loss: 0.0395 - f1_score: 0.9815 - val_loss: 0.0847 - val_f1_score: 0.9786\n",
            "Epoch 46/80\n",
            "128/128 [==============================] - 4s 32ms/step - loss: 0.0333 - f1_score: 0.9794 - val_loss: 0.0888 - val_f1_score: 0.9816\n",
            "Epoch 47/80\n",
            "128/128 [==============================] - 3s 26ms/step - loss: 0.0288 - f1_score: 0.9899 - val_loss: 0.1597 - val_f1_score: 0.9656\n",
            "Epoch 48/80\n",
            "128/128 [==============================] - 4s 30ms/step - loss: 0.0304 - f1_score: 0.9852 - val_loss: 0.0705 - val_f1_score: 0.9735\n",
            "Epoch 49/80\n",
            "128/128 [==============================] - 4s 33ms/step - loss: 0.0256 - f1_score: 0.9829 - val_loss: 0.0558 - val_f1_score: 0.9905\n",
            "Epoch 50/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0306 - f1_score: 0.9863 - val_loss: 0.0569 - val_f1_score: 0.9854\n",
            "Epoch 51/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0399 - f1_score: 0.9771 - val_loss: 0.0730 - val_f1_score: 0.9827\n",
            "Epoch 52/80\n",
            "128/128 [==============================] - 5s 35ms/step - loss: 0.0426 - f1_score: 0.9748 - val_loss: 0.0572 - val_f1_score: 0.9838\n",
            "Epoch 53/80\n",
            "128/128 [==============================] - 3s 26ms/step - loss: 0.0195 - f1_score: 0.9914 - val_loss: 0.0895 - val_f1_score: 0.9816\n",
            "Epoch 54/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0206 - f1_score: 0.9898 - val_loss: 0.0763 - val_f1_score: 0.9783\n",
            "Epoch 55/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0339 - f1_score: 0.9858 - val_loss: 0.0805 - val_f1_score: 0.9854\n",
            "Epoch 56/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0210 - f1_score: 0.9918 - val_loss: 0.0592 - val_f1_score: 0.9818\n",
            "Epoch 57/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0254 - f1_score: 0.9880 - val_loss: 0.1470 - val_f1_score: 0.9712\n",
            "Epoch 58/80\n",
            "128/128 [==============================] - 5s 36ms/step - loss: 0.0269 - f1_score: 0.9889 - val_loss: 0.0546 - val_f1_score: 0.9839\n",
            "Epoch 59/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0262 - f1_score: 0.9879 - val_loss: 0.0754 - val_f1_score: 0.9849\n",
            "Epoch 60/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0205 - f1_score: 0.9908 - val_loss: 0.0808 - val_f1_score: 0.9767\n",
            "Epoch 61/80\n",
            "128/128 [==============================] - 5s 35ms/step - loss: 0.0297 - f1_score: 0.9837 - val_loss: 0.0769 - val_f1_score: 0.9806\n",
            "Epoch 62/80\n",
            "128/128 [==============================] - 3s 26ms/step - loss: 0.0317 - f1_score: 0.9824 - val_loss: 0.0754 - val_f1_score: 0.9769\n",
            "Epoch 63/80\n",
            "128/128 [==============================] - 4s 34ms/step - loss: 0.0390 - f1_score: 0.9881 - val_loss: 0.1715 - val_f1_score: 0.9577\n",
            "Epoch 64/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0261 - f1_score: 0.9802 - val_loss: 0.0831 - val_f1_score: 0.9566\n",
            "Epoch 65/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0230 - f1_score: 0.9928 - val_loss: 0.1012 - val_f1_score: 0.9760\n",
            "Epoch 66/80\n",
            "128/128 [==============================] - 5s 36ms/step - loss: 0.0356 - f1_score: 0.9829 - val_loss: 0.1058 - val_f1_score: 0.9575\n",
            "Epoch 67/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0222 - f1_score: 0.9876 - val_loss: 0.0638 - val_f1_score: 0.9846\n",
            "Epoch 68/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0360 - f1_score: 0.9876 - val_loss: 0.0935 - val_f1_score: 0.9740\n",
            "Epoch 69/80\n",
            "128/128 [==============================] - 5s 36ms/step - loss: 0.0266 - f1_score: 0.9878 - val_loss: 0.0611 - val_f1_score: 0.9868\n",
            "Epoch 70/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0229 - f1_score: 0.9852 - val_loss: 0.1131 - val_f1_score: 0.9723\n",
            "Epoch 71/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0260 - f1_score: 0.9906 - val_loss: 0.0700 - val_f1_score: 0.9846\n",
            "Epoch 72/80\n",
            "128/128 [==============================] - 5s 36ms/step - loss: 0.0391 - f1_score: 0.9789 - val_loss: 0.0714 - val_f1_score: 0.9718\n",
            "Epoch 73/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0203 - f1_score: 0.9884 - val_loss: 0.0477 - val_f1_score: 0.9875\n",
            "Epoch 74/80\n",
            "128/128 [==============================] - 4s 28ms/step - loss: 0.0201 - f1_score: 0.9904 - val_loss: 0.0879 - val_f1_score: 0.9787\n",
            "Epoch 75/80\n",
            "128/128 [==============================] - 5s 37ms/step - loss: 0.0221 - f1_score: 0.9871 - val_loss: 0.0648 - val_f1_score: 0.9848\n",
            "Epoch 76/80\n",
            "128/128 [==============================] - 4s 29ms/step - loss: 0.0289 - f1_score: 0.9880 - val_loss: 0.0625 - val_f1_score: 0.9844\n",
            "Epoch 77/80\n",
            "128/128 [==============================] - 4s 27ms/step - loss: 0.0434 - f1_score: 0.9812 - val_loss: 0.1066 - val_f1_score: 0.9719\n",
            "Epoch 78/80\n",
            "128/128 [==============================] - 4s 33ms/step - loss: 0.0276 - f1_score: 0.9897 - val_loss: 0.0747 - val_f1_score: 0.9833\n",
            "Epoch 79/80\n",
            "128/128 [==============================] - 4s 34ms/step - loss: 0.0155 - f1_score: 0.9913 - val_loss: 0.0578 - val_f1_score: 0.9888\n",
            "Epoch 80/80\n",
            "128/128 [==============================] - 4s 34ms/step - loss: 0.0157 - f1_score: 0.9931 - val_loss: 0.0499 - val_f1_score: 0.9924\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history_plot(history)"
      ],
      "metadata": {
        "id": "-j2bf9lOcJ5-",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "1b5ab7f0-e09d-42e0-bff9-9e61ee71fbf1"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"The best epoch is: {model_checkpoint.best:.3f}\")\n",
        "\n",
        "# Load the model with the best weights\n",
        "best_model_path = '/content/drive/MyDrive/blockchain/best_model_alz_test'\n",
        "best_model = load_model(best_model_path)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iihQadsFEAK0",
        "outputId": "ac36c37b-afa5-46d8-f8a6-7eecc99b21c3"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The best epoch is: 0.992\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_true, y_pred = predictions(test_dataset, best_model)\n",
        "# Generate and print the classification report\n",
        "class_names = LABELS_ALZ if train_path == DATASET_TRAIN_PATH_ALZ else LABELS_TUMOR\n",
        "print(\"\\t\\tClassification report:\\n\")\n",
        "print(classification_report(y_true, y_pred,target_names=class_names))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "diyxDxdZETMd",
        "outputId": "96e314ea-b82d-43f9-96ee-9a886fe29896"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\t\tClassification report:\n",
            "\n",
            "                  precision    recall  f1-score   support\n",
            "\n",
            "     NonDemented       0.99      0.99      0.99       512\n",
            "VeryMildDemented       0.99      0.99      0.99       359\n",
            "    MildDemented       0.99      0.99      0.99       144\n",
            "ModerateDemented       1.00      1.00      1.00        10\n",
            "\n",
            "        accuracy                           0.99      1025\n",
            "       macro avg       0.99      0.99      0.99      1025\n",
            "    weighted avg       0.99      0.99      0.99      1025\n",
            "\n"
          ]
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}