MIO_TCD_CNN.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Pdaret/MIO-TCD-CNN/blob/main/MIO_TCD_CNN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Extracting file \n",
        "we use **dataset** rar file an extract it to colab runtime `sample_data` folder"
      ],
      "metadata": {
        "id": "tE7CuOwvpSlr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive', force_remount=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uggr_-uhm0ct",
        "outputId": "acfc88a9-8d77-43d6-bf0b-843de7f1784a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!unrar x \"/content/drive/MyDrive/HW02-Practical.rar\" \"/content/sample_data/miotcd\""
      ],
      "metadata": {
        "id": "RqxHbwmInPL_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "optional"
      ],
      "metadata": {
        "id": "NiRGM9zwpegc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!apt\n",
        "from pyunpack import Archive\n",
        "Archive('/content/drive/MyDrive/HW02-Practical.rar').extractall('/content/sample_data/miotcd')"
      ],
      "metadata": {
        "id": "4YEDRvz8xbna"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eQLpwxROLMjy"
      },
      "source": [
        "# <font color = 'red'># Part 1</font>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bei6FmA1-80Q"
      },
      "source": [
        "# Under Fitting Model \n",
        "\n",
        "To induce underfitting in our model, we can make it less complex or reduce the capacity of the model. Here are a few modifications we can make to our model to encourage underfitting:\n",
        "\n",
        "1. Decrease the number of filters in the convolutional layers: By reducing the number of filters, the model will extract fewer features, leading to a simpler representation.\n",
        "\n",
        "2. Remove a convolutional layer: Removing a layer reduces the depth of the model and further simplifies it.\n",
        "\n",
        "3. Reduce the number of neurons in the dense layer: By decreasing the number of neurons, you reduce the complexity of the model's fully connected layers."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vEzou18eC9Nv",
        "outputId": "559afd75-2109-43b4-dfa6-0ce603d02fcf"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 15000 images belonging to 5 classes.\n",
            "Found 5000 images belonging to 5 classes.\n",
            "Found 5000 images belonging to 5 classes.\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "train_datagen = ImageDataGenerator(rescale=1./255)\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    '/content/sample_data/miotcd/HW02-Practical/Dataset/train',\n",
        "    target_size=(128, 128),\n",
        "    batch_size=128,\n",
        "    class_mode='categorical'\n",
        ")\n",
        "\n",
        "validation_datagen = ImageDataGenerator(rescale=1./255)\n",
        "validation_generator = validation_datagen.flow_from_directory(\n",
        "    '/content/sample_data/miotcd/HW02-Practical/Dataset/valid',\n",
        "    target_size=(128, 128),\n",
        "    batch_size=128,\n",
        "    class_mode='categorical'\n",
        ")\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "test_generator = test_datagen.flow_from_directory(\n",
        "        '/content/sample_data/miotcd/HW02-Practical/Dataset/test',\n",
        "        target_size=(128, 128),\n",
        "        batch_size=512,\n",
        "        class_mode='categorical'\n",
        ")\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HZjy9Qs8JCPm"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n",
        "\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Conv2D(16, (3, 3), activation='relu', input_shape=(128, 128, 3)))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(32, activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(5, activation='softmax'))\n",
        "\n",
        "\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VrnQrR7nm7Xc",
        "outputId": "b241c15e-f822-4ece-84ab-4b51c7e1f12b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "117/117 [==============================] - 40s 253ms/step - loss: 1.8079 - accuracy: 0.1963 - val_loss: 1.6094 - val_accuracy: 0.1999\n",
            "Epoch 2/10\n",
            "117/117 [==============================] - 30s 255ms/step - loss: 1.6095 - accuracy: 0.1934 - val_loss: 1.6094 - val_accuracy: 0.1995\n",
            "Epoch 3/10\n",
            "117/117 [==============================] - 31s 262ms/step - loss: 1.6095 - accuracy: 0.1952 - val_loss: 1.6094 - val_accuracy: 0.2001\n",
            "Epoch 4/10\n",
            "117/117 [==============================] - 29s 247ms/step - loss: 1.6095 - accuracy: 0.1898 - val_loss: 1.6094 - val_accuracy: 0.1999\n",
            "Epoch 5/10\n",
            "117/117 [==============================] - 31s 269ms/step - loss: 1.6095 - accuracy: 0.2002 - val_loss: 1.6094 - val_accuracy: 0.1997\n",
            "Epoch 6/10\n",
            "117/117 [==============================] - 29s 247ms/step - loss: 1.6095 - accuracy: 0.1974 - val_loss: 1.6094 - val_accuracy: 0.2001\n",
            "Epoch 7/10\n",
            "117/117 [==============================] - 29s 246ms/step - loss: 1.6095 - accuracy: 0.1958 - val_loss: 1.6094 - val_accuracy: 0.2001\n",
            "Epoch 8/10\n",
            "117/117 [==============================] - 25s 218ms/step - loss: 1.6095 - accuracy: 0.1951 - val_loss: 1.6094 - val_accuracy: 0.1997\n",
            "Epoch 9/10\n",
            "117/117 [==============================] - 30s 255ms/step - loss: 1.6095 - accuracy: 0.1915 - val_loss: 1.6094 - val_accuracy: 0.1999\n",
            "Epoch 10/10\n",
            "117/117 [==============================] - 30s 255ms/step - loss: 1.6095 - accuracy: 0.1969 - val_loss: 1.6094 - val_accuracy: 0.2001\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(\n",
        "    train_generator,\n",
        "    steps_per_epoch=train_generator.n // train_generator.batch_size,\n",
        "    epochs=10,\n",
        "    validation_data=validation_generator,\n",
        "    validation_steps=validation_generator.n // validation_generator.batch_size\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "bwzcLLmb-SOO",
        "outputId": "2115a55a-2230-444a-b395-2342526d8dc2"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAHHCAYAAABXx+fLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABQa0lEQVR4nO3deVxU9f4/8NcBZNgHQUFQFkXcFUnUq1yXilQsbmQuKRkureKW6Ve95l5aZuZSmRaKlWbXBbJUDM0lzVJTzNIUAhUF3FlFUObz+2N+c3IClGXgzPJ6Ph7zYObM55zzPswUL895n3MkIYQAERERkQWxUroAIiIiorrGAEREREQWhwGIiIiILA4DEBEREVkcBiAiIiKyOAxAREREZHEYgIiIiMjiMAARERGRxWEAIiIiIovDAERkZEaMGAF/f/9qzTtnzhxIkmTYgozM+fPnIUkS4uLi6nzdkiRhzpw58uu4uDhIkoTz588/dF5/f3+MGDHCoPXU5LtCZOkYgIgqSZKkSj327dundKkWb/z48ZAkCampqRWOmTFjBiRJwm+//VaHlVVdZmYm5syZg+TkZKVLkelC6OLFi5UuhajabJQugMhUfPHFF3qvP//8cyQlJZWZ3rp16xqt59NPP4VGo6nWvG+++SamTZtWo/Wbg6ioKKxYsQIbNmzArFmzyh3z1VdfoX379ujQoUO11zN8+HA899xzUKlU1V7Gw2RmZmLu3Lnw9/dHx44d9d6ryXeFyNIxABFV0vPPP6/3+ueff0ZSUlKZ6f90+/ZtODg4VHo99erVq1Z9AGBjYwMbG/5n3bVrVzRv3hxfffVVuQHo8OHDSE9PxzvvvFOj9VhbW8Pa2rpGy6iJmnxXiCwdD4ERGVDv3r3Rrl07/Prrr+jZsyccHBzw3//+FwDwzTff4Mknn4S3tzdUKhUCAgIwf/58lJaW6i3jn30d9x9uWL16NQICAqBSqdC5c2ccPXpUb97yeoAkScLYsWORkJCAdu3aQaVSoW3btkhMTCxT/759+xASEgI7OzsEBARg1apVle4r+vHHHzFo0CD4+vpCpVLBx8cHr7/+OoqKispsn5OTEy5fvozIyEg4OTmhYcOGmDx5cpnfRU5ODkaMGAG1Wg1XV1dER0cjJyfnobUA2r1Af/75J44fP17mvQ0bNkCSJAwdOhQlJSWYNWsWOnXqBLVaDUdHR/To0QN79+596DrK6wESQuCtt95CkyZN4ODggEcffRR//PFHmXlv3ryJyZMno3379nBycoKLiwvCw8Nx8uRJecy+ffvQuXNnAMDIkSPlw6y6/qfyeoAKCwvxxhtvwMfHByqVCi1btsTixYshhNAbV5XvRXVdvXoVo0ePhqenJ+zs7BAUFIR169aVGbdx40Z06tQJzs7OcHFxQfv27bFs2TL5/bt372Lu3LkIDAyEnZ0d3N3d8e9//xtJSUkGq5UsD/+pSGRgN27cQHh4OJ577jk8//zz8PT0BKD9Y+nk5IRJkybByckJP/zwA2bNmoW8vDy89957D13uhg0bkJ+fj1deeQWSJGHRokUYMGAA0tLSHron4ODBg9i6dSvGjBkDZ2dnLF++HM8++ywuXrwId3d3AMCJEyfQr18/eHl5Ye7cuSgtLcW8efPQsGHDSm33pk2bcPv2bbz22mtwd3fHkSNHsGLFCly6dAmbNm3SG1taWoq+ffuia9euWLx4MXbv3o33338fAQEBeO211wBog8TTTz+NgwcP4tVXX0Xr1q0RHx+P6OjoStUTFRWFuXPnYsOGDXjkkUf01v2///0PPXr0gK+vL65fv47PPvsMQ4cOxUsvvYT8/HzExsaib9++OHLkSJnDTg8za9YsvPXWW+jfvz/69++P48ePo0+fPigpKdEbl5aWhoSEBAwaNAhNmzbFlStXsGrVKvTq1QunT5+Gt7c3WrdujXnz5mHWrFl4+eWX0aNHDwBA9+7dy123EAL/+c9/sHfvXowePRodO3bErl27MGXKFFy+fBkffPCB3vjKfC+qq6ioCL1790ZqairGjh2Lpk2bYtOmTRgxYgRycnIwYcIEAEBSUhKGDh2Kxx9/HO+++y4A4MyZMzh06JA8Zs6cOVi4cCFefPFFdOnSBXl5eTh27BiOHz+OJ554okZ1kgUTRFQtMTEx4p//CfXq1UsAEJ988kmZ8bdv3y4z7ZVXXhEODg7izp078rTo6Gjh5+cnv05PTxcAhLu7u7h586Y8/ZtvvhEAxLfffitPmz17dpmaAAhbW1uRmpoqTzt58qQAIFasWCFPi4iIEA4ODuLy5cvytJSUFGFjY1NmmeUpb/sWLlwoJEkSFy5c0Ns+AGLevHl6Y4ODg0WnTp3k1wkJCQKAWLRokTzt3r17okePHgKAWLt27UNr6ty5s2jSpIkoLS2VpyUmJgoAYtWqVfIyi4uL9ea7deuW8PT0FKNGjdKbDkDMnj1bfr127VoBQKSnpwshhLh69aqwtbUVTz75pNBoNPK4//73vwKAiI6OlqfduXNHry4htJ+1SqXS+90cPXq0wu3953dF9zt766239MYNHDhQSJKk9x2o7PeiPLrv5HvvvVfhmKVLlwoA4ssvv5SnlZSUiG7dugknJyeRl5cnhBBiwoQJwsXFRdy7d6/CZQUFBYknn3zygTURVRUPgREZmEqlwsiRI8tMt7e3l5/n5+fj+vXr6NGjB27fvo0///zzocsdMmQI6tevL7/W7Q1IS0t76LxhYWEICAiQX3fo0AEuLi7yvKWlpdi9ezciIyPh7e0tj2vevDnCw8MfunxAf/sKCwtx/fp1dO/eHUIInDhxosz4V199Ve91jx499LZlx44dsLGxkfcIAdqem3HjxlWqHkDbt3Xp0iUcOHBAnrZhwwbY2tpi0KBB8jJtbW0BABqNBjdv3sS9e/cQEhJS7uGzB9m9ezdKSkowbtw4vcOGEydOLDNWpVLBykr7v+DS0lLcuHEDTk5OaNmyZZXXq7Njxw5YW1tj/PjxetPfeOMNCCGwc+dOvekP+17UxI4dO9CoUSMMHTpUnlavXj2MHz8eBQUF2L9/PwDA1dUVhYWFDzyc5erqij/++AMpKSk1rotIhwGIyMAaN24s/0G93x9//IFnnnkGarUaLi4uaNiwodxAnZub+9Dl+vr66r3WhaFbt25VeV7d/Lp5r169iqKiIjRv3rzMuPKmlefixYsYMWIE3Nzc5L6eXr16ASi7fXZ2dmUOrd1fDwBcuHABXl5ecHJy0hvXsmXLStUDAM899xysra2xYcMGAMCdO3cQHx+P8PBwvTC5bt06dOjQQe4vadiwIbZv316pz+V+Fy5cAAAEBgbqTW/YsKHe+gBt2Prggw8QGBgIlUqFBg0aoGHDhvjtt9+qvN771+/t7Q1nZ2e96bozE3X16Tzse1ETFy5cQGBgoBzyKqplzJgxaNGiBcLDw9GkSROMGjWqTB/SvHnzkJOTgxYtWqB9+/aYMmWK0V++gIwfAxCRgd2/J0QnJycHvXr1wsmTJzFv3jx8++23SEpKknseKnMqc0VnG4l/NLcaet7KKC0txRNPPIHt27dj6tSpSEhIQFJSktys+8/tq6szpzw8PPDEE09gy5YtuHv3Lr799lvk5+cjKipKHvPll19ixIgRCAgIQGxsLBITE5GUlITHHnusVk8xX7BgASZNmoSePXviyy+/xK5du5CUlIS2bdvW2anttf29qAwPDw8kJydj27Ztcv9SeHi4Xq9Xz5498ddff2HNmjVo164dPvvsMzzyyCP47LPP6qxOMj9sgiaqA/v27cONGzewdetW9OzZU56enp6uYFV/8/DwgJ2dXbkXDnzQxQR1Tp06hXPnzmHdunV44YUX5Ok1OUvHz88Pe/bsQUFBgd5eoLNnz1ZpOVFRUUhMTMTOnTuxYcMGuLi4ICIiQn5/8+bNaNasGbZu3ap32Gr27NnVqhkAUlJS0KxZM3n6tWvXyuxV2bx5Mx599FHExsbqTc/JyUGDBg3k11W5srefnx92796N/Px8vb1AukOsuvrqgp+fH3777TdoNBq9vUDl1WJra4uIiAhERERAo9FgzJgxWLVqFWbOnCnvgXRzc8PIkSMxcuRIFBQUoGfPnpgzZw5efPHFOtsmMi/cA0RUB3T/0r7/X9YlJSX4+OOPlSpJj7W1NcLCwpCQkIDMzEx5empqapm+kYrmB/S3TwihdypzVfXv3x/37t3DypUr5WmlpaVYsWJFlZYTGRkJBwcHfPzxx9i5cycGDBgAOzu7B9b+yy+/4PDhw1WuOSwsDPXq1cOKFSv0lrd06dIyY62trcvsadm0aRMuX76sN83R0REAKnX6f//+/VFaWooPP/xQb/oHH3wASZIq3c9lCP3790d2dja+/vpredq9e/ewYsUKODk5yYdHb9y4oTeflZWVfHHK4uLicsc4OTmhefPm8vtE1cE9QER1oHv37qhfvz6io6Pl2zR88cUXdXqo4WHmzJmD77//HqGhoXjttdfkP6Tt2rV76G0YWrVqhYCAAEyePBmXL1+Gi4sLtmzZUqNekoiICISGhmLatGk4f/482rRpg61bt1a5P8bJyQmRkZFyH9D9h78A4KmnnsLWrVvxzDPP4Mknn0R6ejo++eQTtGnTBgUFBVVal+56RgsXLsRTTz2F/v3748SJE9i5c6feXh3deufNm4eRI0eie/fuOHXqFNavX6+35wgAAgIC4Orqik8++QTOzs5wdHRE165d0bRp0zLrj4iIwKOPPooZM2bg/PnzCAoKwvfff49vvvkGEydO1Gt4NoQ9e/bgzp07ZaZHRkbi5ZdfxqpVqzBixAj8+uuv8Pf3x+bNm3Ho0CEsXbpU3kP14osv4ubNm3jsscfQpEkTXLhwAStWrEDHjh3lfqE2bdqgd+/e6NSpE9zc3HDs2DFs3rwZY8eONej2kIVR5uQzItNX0Wnwbdu2LXf8oUOHxL/+9S9hb28vvL29xf/93/+JXbt2CQBi79698riKToMv75Rj/OO07IpOg4+JiSkzr5+fn95p2UIIsWfPHhEcHCxsbW1FQECA+Oyzz8Qbb7wh7OzsKvgt/O306dMiLCxMODk5iQYNGoiXXnpJPq36/lO4o6OjhaOjY5n5y6v9xo0bYvjw4cLFxUWo1WoxfPhwceLEiUqfBq+zfft2AUB4eXmVOfVco9GIBQsWCD8/P6FSqURwcLD47rvvynwOQjz8NHghhCgtLRVz584VXl5ewt7eXvTu3Vv8/vvvZX7fd+7cEW+88YY8LjQ0VBw+fFj06tVL9OrVS2+933zzjWjTpo18SQLdtpdXY35+vnj99deFt7e3qFevnggMDBTvvfee3mn5um2p7Pfin3TfyYoeX3zxhRBCiCtXroiRI0eKBg0aCFtbW9G+ffsyn9vmzZtFnz59hIeHh7C1tRW+vr7ilVdeEVlZWfKYt956S3Tp0kW4uroKe3t70apVK/H222+LkpKSB9ZJ9CCSEEb0T1AiMjqRkZE8BZmIzA57gIhI9s/bVqSkpGDHjh3o3bu3MgUREdUS7gEiIpmXlxdGjBiBZs2a4cKFC1i5ciWKi4tx4sSJMte2ISIyZWyCJiJZv3798NVXXyE7OxsqlQrdunXDggULGH6IyOxwDxARERFZHPYAERERkcVhACIiIiKLwx6gcmg0GmRmZsLZ2blKl6EnIiIi5QghkJ+fD29v7zI34v0nBqByZGZmwsfHR+kyiIiIqBoyMjLQpEmTB45hACqH7hLtGRkZcHFxUbgaIiIiqoy8vDz4+Pjo3Qy4IgxA5dAd9nJxcWEAIiIiMjGVaV9hEzQRERFZHEUD0IEDBxAREQFvb29IkoSEhIQHjh8xYgQkSSrzaNu2bbnj33nnHUiShIkTJxq+eCIiIjJZigagwsJCBAUF4aOPPqrU+GXLliErK0t+ZGRkwM3NDYMGDSoz9ujRo1i1ahU6dOhg6LKJiIjIxCnaAxQeHo7w8PBKj1er1VCr1fLrhIQE3Lp1CyNHjtQbV1BQgKioKHz66ad46623DFYvERFVT2lpKe7evat0GWQGbG1tH3qKe2WYdBN0bGwswsLC4Ofnpzc9JiYGTz75JMLCwioVgIqLi1FcXCy/zsvLM3itRESWSAiB7Oxs5OTkKF0KmQkrKys0bdoUtra2NVqOyQagzMxM7Ny5Exs2bNCbvnHjRhw/fhxHjx6t9LIWLlyIuXPnGrpEIiKLpws/Hh4ecHBw4MVlqUZ0FyrOysqCr69vjb5PJhuA1q1bB1dXV0RGRsrTMjIyMGHCBCQlJcHOzq7Sy5o+fTomTZokv9ZdR4CIiKqvtLRUDj/u7u5Kl0NmomHDhsjMzMS9e/dQr169ai/HJAOQEAJr1qzB8OHD9XaB/frrr7h69SoeeeQReVppaSkOHDiADz/8EMXFxbC2ti6zPJVKBZVKVSe1ExFZCl3Pj4ODg8KVkDnR/d0vLS21vAC0f/9+pKamYvTo0XrTH3/8cZw6dUpv2siRI9GqVStMnTq13PBDRES1i4e9yJAM9X1SNAAVFBQgNTVVfp2eno7k5GS4ubnB19cX06dPx+XLl/H555/rzRcbG4uuXbuiXbt2etOdnZ3LTHN0dIS7u3uZ6URERGS5FL0O0LFjxxAcHIzg4GAAwKRJkxAcHIxZs2YBALKysnDx4kW9eXJzc7Fly5Yye3+IiIiMmb+/P5YuXVrp8fv27YMkSbV+Bl1cXBxcXV1rdR3GSNE9QL1794YQosL34+LiykxTq9W4fft2pdexb9++alRGRESW6mGHWGbPno05c+ZUeblHjx6Fo6Njpcd3794dWVlZete/I8MxyR4gUyUEcO0acOsW0LKl0tUQEVF5srKy5Odff/01Zs2ahbNnz8rTnJyc5OdCCJSWlsLG5uF/Ths2bFilOmxtbdGoUaMqzUOVx5uh1qHERMDTExgyROlKiIioIo0aNZIfarUakiTJr//88084Oztj586d6NSpE1QqFQ4ePIi//voLTz/9NDw9PeHk5ITOnTtj9+7desv95yEwSZLw2Wef4ZlnnoGDgwMCAwOxbds2+f1/HgLTHaratWsXWrduDScnJ/Tr108vsN27dw/jx4+Hq6sr3N3dMXXqVERHR+tdMqYyVq5ciYCAANja2qJly5b44osv5PeEEJgzZw58fX2hUqng7e2N8ePHy+9//PHHCAwMhJ2dHTw9PTFw4MAqrbuuMADVoebNtT9TUgCNRtlaiIiUIARQWKjM4wEdF1U2bdo0vPPOOzhz5gw6dOiAgoIC9O/fH3v27MGJEyfQr18/RERElOlj/ae5c+di8ODB+O2339C/f39ERUXh5s2bFY6/ffs2Fi9ejC+++AIHDhzAxYsXMXnyZPn9d999F+vXr8fatWtx6NAh5OXlPfRG4/8UHx+PCRMm4I033sDvv/+OV155BSNHjsTevXsBAFu2bMEHH3yAVatWISUlBQkJCWjfvj0AbW/v+PHjMW/ePJw9exaJiYno2bNnldZfZwSVkZubKwCI3Nxcgy63pEQIGxshACEyMgy6aCIio1NUVCROnz4tioqK5GkFBdr/ByrxKCio+jasXbtWqNVq+fXevXsFAJGQkPDQedu2bStWrFghv/bz8xMffPCB/BqAePPNN+/73RQIAGLnzp1667p165ZcCwCRmpoqz/PRRx8JT09P+bWnp6d477335Nf37t0Tvr6+4umnn670Nnbv3l289NJLemMGDRok+vfvL4QQ4v333xctWrQQJSUlZZa1ZcsW4eLiIvLy8ipcX02V973Sqcrfb+4BqkP16gHNmmmfnzunbC1ERFR9ISEheq8LCgowefJktG7dGq6urnBycsKZM2ceugeoQ4cO8nNHR0e4uLjg6tWrFY53cHBAQECA/NrLy0sen5ubiytXrqBLly7y+9bW1ujUqVOVtu3MmTMIDQ3VmxYaGoozZ84AAAYNGoSioiI0a9YML730EuLj43Hv3j0AwBNPPAE/Pz80a9YMw4cPx/r166t04lJdYgCqYy1aaH8yABGRJXJwAAoKlHkY8oLU/zyba/LkyYiPj8eCBQvw448/Ijk5Ge3bt0dJSckDl/PPKxlLkgTNA3okyhsvDHlsrxJ8fHxw9uxZfPzxx7C3t8eYMWPQs2dP3L17F87Ozjh+/Di++uoreHl5YdasWQgKCjLKm+EyANUx3dlf951QQERkMSQJcHRU5lGbF6Q+dOgQRowYgWeeeQbt27dHo0aNcP78+dpbYTnUajU8PT31bgZeWlqK48ePV2k5rVu3xqFDh/SmHTp0CG3atJFf29vbIyIiAsuXL8e+fftw+PBh+U4MNjY2CAsLw6JFi/Dbb7/h/Pnz+OGHH2qwZbWDp8HXMe4BIiIyP4GBgdi6dSsiIiIgSRJmzpz5wD05tWXcuHFYuHAhmjdvjlatWmHFihW4detWlW4fMWXKFAwePBjBwcEICwvDt99+i61bt8pntcXFxaG0tBRdu3aFg4MDvvzyS9jb28PPzw/fffcd0tLS0LNnT9SvXx87duyARqNBSyO89gsDUB1jACIiMj9LlizBqFGj0L17dzRo0ABTp05FXl5endcxdepUZGdn44UXXoC1tTVefvll9O3bt0r3woyMjMSyZcuwePFiTJgwAU2bNsXatWvRu3dvAICrqyveeecdTJo0CaWlpWjfvj2+/fZbuLu7w9XVFVu3bsWcOXNw584dBAYG4quvvkLbtm1raYurTxJ1ffDQBOTl5UGtViM3NxcuLi4GXXZmJtC4MWBtDdy+Ddx3M3siIrNy584dpKeno2nTprCzs1O6HIuk0WjQunVrDB48GPPnz1e6HIN40PeqKn+/uQeojnl5AU5O2oa89HReEZqIiAznwoUL+P7779GrVy8UFxfjww8/RHp6OoYNG6Z0aUaHTdB1TJL+PgzGRmgiIjIkKysrxMXFoXPnzggNDcWpU6ewe/dutG7dWunSjA73ACmgRQvg+HH2ARERkWH5+PiUOYOLysc9QApgIzQREZGyGIAUwABERESkLAYgBTAAERERKYsBSAG6AJSVBShwmQgiIiKLxwCkALUa8PTUPk9JUbYWIiIiS8QApBAeBiMiIlIOA5BCGICIiMxb7969MXHiRPm1v78/li5d+sB5JElCQkJCjddtqOU8yJw5c9CxY8daXUdtYgBSCAMQEZFxioiIQL9+/cp978cff4QkSfjtt9+qvNyjR4/i5Zdfrml5eioKIVlZWQgPDzfouswNA5BCdLfA4NWgiYiMy+jRo5GUlIRLly6VeW/t2rUICQlBhw4dqrzchg0bwsHBwRAlPlSjRo2gUqnqZF2migFIIffvAeLtaImIjMdTTz2Fhg0bIi4uTm96QUEBNm3ahNGjR+PGjRsYOnQoGjduDAcHB7Rv3x5fffXVA5f7z0NgKSkp6NmzJ+zs7NCmTRskJSWVmWfq1Klo0aIFHBwc0KxZM8ycORN3794FAMTFxWHu3Lk4efIkJEmCJElyzf88BHbq1Ck89thjsLe3h7u7O15++WUUFBTI748YMQKRkZFYvHgxvLy84O7ujpiYGHldlaHRaDBv3jw0adIEKpUKHTt2RGJiovx+SUkJxo4dCy8vL9jZ2cHPzw8LFy4EAAghMGfOHPj6+kKlUsHb2xvjx4+v9Lqrg7fCUEizZoCVFZCfD1y5AjRqpHRFRER1QAjg9m1l1u3goL0h40PY2NjghRdeQFxcHGbMmAHp/8+zadMmlJaWYujQoSgoKECnTp0wdepUuLi4YPv27Rg+fDgCAgLQpUuXh65Do9FgwIAB8PT0xC+//ILc3Fy9fiEdZ2dnxMXFwdvbG6dOncJLL70EZ2dn/N///R+GDBmC33//HYmJidi9ezcAQK1Wl1lGYWEh+vbti27duuHo0aO4evUqXnzxRYwdO1Yv5O3duxdeXl7Yu3cvUlNTMWTIEHTs2BEvvfTSQ7cHAJYtW4b3338fq1atQnBwMNasWYP//Oc/+OOPPxAYGIjly5dj27Zt+N///gdfX19kZGQgIyMDALBlyxZ88MEH2LhxI9q2bYvs7GycPHmyUuutNkFl5ObmCgAiNze3VtfTrJkQgBD799fqaoiIFFFUVCROnz4tioqK/p5YUKD9H58Sj4KCStd+5swZAUDs3btXntajRw/x/PPPVzjPk08+Kd544w35da9evcSECRPk135+fuKDDz4QQgixa9cuYWNjIy5fviy/v3PnTgFAxMfHV7iO9957T3Tq1El+PXv2bBEUFFRm3P3LWb16tahfv74ouG/7t2/fLqysrER2drYQQojo6Gjh5+cn7t27J48ZNGiQGDJkSIW1/HPd3t7e4u2339Yb07lzZzFmzBghhBDjxo0Tjz32mNBoNGWW9f7774sWLVqIkpKSCtenU+736v+ryt9vHgJTEBuhiYiMU6tWrdC9e3esWbMGAJCamooff/wRo0ePBgCUlpZi/vz5aN++Pdzc3ODk5IRdu3bh4sWLlVr+mTNn4OPjA29vb3lat27dyoz7+uuvERoaikaNGsHJyQlvvvlmpddx/7qCgoLg6OgoTwsNDYVGo8HZ+xpR27ZtC2tra/m1l5cXrl69Wql15OXlITMzE6GhoXrTQ0NDcebMGQDaw2zJyclo2bIlxo8fj++//14eN2jQIBQVFaFZs2Z46aWXEB8fj3v37lVpO6uKAUhBbIQmIovj4AAUFCjzqGID8ujRo7Flyxbk5+dj7dq1CAgIQK9evQAA7733HpYtW4apU6di7969SE5ORt++fVFSUmKwX9Xhw4cRFRWF/v3747vvvsOJEycwY8YMg67jfvXq1dN7LUkSNBqNwZb/yCOPID09HfPnz0dRUREGDx6MgQMHAtDexf7s2bP4+OOPYW9vjzFjxqBnz55V6kGqKvYAKYh7gIjI4kgScN+eCGM2ePBgTJgwARs2bMDnn3+O1157Te4HOnToEJ5++mk8//zzALQ9PefOnUObNm0qtezWrVsjIyMDWVlZ8PLyAgD8/PPPemN++ukn+Pn5YcaMGfK0Cxcu6I2xtbVFaWnpQ9cVFxeHwsJCeS/QoUOHYGVlhZa6f4nXkIuLC7y9vXHo0CE5JOrWc39PlIuLC4YMGYIhQ4Zg4MCB6NevH27evAk3NzfY29sjIiICERERiImJQatWrXDq1Ck88sgjBqnxnxiAFMQARERkvJycnDBkyBBMnz4deXl5GDFihPxeYGAgNm/ejJ9++gn169fHkiVLcOXKlUoHoLCwMLRo0QLR0dF47733kJeXpxd0dOu4ePEiNm7ciM6dO2P79u2Ij4/XG+Pv74/09HQkJyejSZMmcHZ2LnP6e1RUFGbPno3o6GjMmTMH165dw7hx4zB8+HB46u7LZABTpkzB7NmzERAQgI4dO2Lt2rVITk7G+vXrAQBLliyBl5cXgoODYWVlhU2bNqFRo0ZwdXVFXFwcSktL0bVrVzg4OODLL7+Evb09/Pz8DFbfP/EQmIJ0Aeivv4BaPtRJRETVMHr0aNy6dQt9+/bV69d588038cgjj6Bv377o3bs3GjVqhMjIyEov18rKCvHx8SgqKkKXLl3w4osv4u2339Yb85///Aevv/46xo4di44dO+Knn37CzJkz9cY8++yz6NevHx599FE0bNiw3FPxHRwcsGvXLty8eROdO3fGwIED8fjjj+PDDz+s2i/jIcaPH49JkybhjTfeQPv27ZGYmIht27YhMDAQgPaMtkWLFiEkJASdO3fG+fPnsWPHDlhZWcHV1RWffvopQkND0aFDB+zevRvffvst3N3dDVrj/SQheBWaf8rLy4NarUZubi5cXFxqbT0aDeDkBBQVAampQEBAra2KiKjO3blzB+np6WjatCns7OyULofMxIO+V1X5+809QAqysgL+fzBmIzQREVEdYgBSGPuAiIiI6h4DkMIYgIiIiOoeA5DCGICIiIjqHgOQwhiAiMjc8VwbMiRDfZ8YgBSmuwZVRoZy9wckIqoNuisL3+b/3MiAdFfCvv+2HdXBCyEqzM0NcHcHbtwAUlKAoCClKyIiMgxra2u4urrK95NycHCQr6RMVB0ajQbXrl2Dg4MDbGxqFmEYgIxAixbA4cPaw2AMQERkTho1agQAlb6pJtHDWFlZwdfXt8ZhmgHICNwfgIiIzIkkSfDy8oKHh0et3tiSLIetrS2srGrewcMAZATYCE1E5s7a2rrGPRtEhsQmaCPAAERERFS3FA1ABw4cQEREBLy9vSFJEhISEh44fsSIEZAkqcyjbdu28piFCxeic+fOcHZ2hoeHByIjI3HWyO8zoTsT7OxZgGeLEhER1T5FA1BhYSGCgoLw0UcfVWr8smXLkJWVJT8yMjLg5uaGQYMGyWP279+PmJgY/Pzzz0hKSsLdu3fRp08fFBYW1tZm1Fjz5tqft25pzwYjIiKi2qVoD1B4eDjCw8MrPV6tVkOtVsuvExIScOvWLYwcOVKelpiYqDdPXFwcPDw88Ouvv6Jnz541L7oW2NsDvr7AxYvaw2ANGihdERERkXkz6R6g2NhYhIWFwc/Pr8Ixubm5AAA3N7cKxxQXFyMvL0/vUdfYB0RERFR3TDYAZWZmYufOnXjxxRcrHKPRaDBx4kSEhoaiXbt2FY5buHChvHdJrVbDx8enNkp+IAYgIiKiumOyAWjdunVwdXVFZGRkhWNiYmLw+++/Y+PGjQ9c1vTp05Gbmys/MjIyDFztw93fCE1ERES1yySvAySEwJo1azB8+HDY2tqWO2bs2LH47rvvcODAATRp0uSBy1OpVFCpVLVRaqVxDxAREVHdMck9QPv370dqaipGjx5d5j0hBMaOHYv4+Hj88MMPaNq0qQIVVp0uAKWkABqNsrUQERGZO0X3ABUUFCA1NVV+nZ6ejuTkZLi5ucHX1xfTp0/H5cuX8fnnn+vNFxsbi65du5bb1xMTE4MNGzbgm2++gbOzM7KzswFozyCzt7ev3Q2qAT8/oF49oLhYe2f4B/R1ExERUQ0pugfo2LFjCA4ORnBwMABg0qRJCA4OxqxZswAAWVlZuHjxot48ubm52LJlS7l7fwBg5cqVyM3NRe/eveHl5SU/vv7669rdmBqytv77ekA8DEZERFS7FN0D1Lt3b4gHXPo4Li6uzDS1Wo3bt29XOM+DlmfsWrYEzpzRNkI/8YTS1RAREZkvk+wBMldshCYiIqobDEBGhAGIiIiobjAAGREGICIiorrBAGREdAHo/Hnt2WBERERUOxiAjIiHB6BWA0IA910dgIiIiAyMAciISBIPgxEREdUFBiAjwwBERERU+xiAjAwDEBERUe1jADIyDEBERES1jwHIyLRsqf159qyydRAREZkzBiAjExio/XntGnDrlrK1EBERmSsGICPj5AR4e2ufp6QoWwsREZG5YgAyQuwDIiIiql0MQEaIAYiIiKh2MQAZIV0jNAMQERFR7WAAMkK6PUA8E4yIiKh2MAAZofsPgQmhbC1ERETmiAHICDVtClhbA7dvA5mZSldDRERkfhiAjFC9ekCzZtrn7AMiIiIyPAYgI8UzwYiIiGoPA5CR4i0xiIiIag8DkJHiHiAiIqLawwBkpBiAiIiIag8DkJHSBaC0NODuXWVrISIiMjcMQEbK2xtwdARKS4H0dKWrISIiMi8MQEZKknhFaCIiotrCAGTE2AdERERUOxiAjBgDEBERUe1gADJiDEBERES1gwHIiDEAERER1Q4GICOmC0CZmUB+vrK1EBERmRMGICPm6gp4eGifp6QoWgoREZFZYQAycjwMRkREZHgMQEaOAYiIiMjwGICMHAMQERGR4TEAGbmWLbU/eTVoIiIiw2EAMnL37wESQtlaiIiIzAUDkJELCNDeFywvD7h6VelqiIiIzAMDkJFTqQB/f+1z9gEREREZBgOQCWAjNBERkWExAJkANkITEREZFgOQCeAeICIiIsNSNAAdOHAAERER8Pb2hiRJSEhIeOD4ESNGQJKkMo+2bdvqjfvoo4/g7+8POzs7dO3aFUeOHKnFrah9DEBERESGpWgAKiwsRFBQED766KNKjV+2bBmysrLkR0ZGBtzc3DBo0CB5zNdff41JkyZh9uzZOH78OIKCgtC3b19cNeFTqHQBKDUVKC1VthYiIiJzIAlhHFeXkSQJ8fHxiIyMrPQ8CQkJGDBgANLT0+Hn5wcA6Nq1Kzp37owPP/wQAKDRaODj44Nx48Zh2rRplVpuXl4e1Go1cnNz4eLiUuVtMTSNBnB0BO7cAf76C2jWTOmKiIiIjE9V/n6bdA9QbGwswsLC5PBTUlKCX3/9FWFhYfIYKysrhIWF4fDhwxUup7i4GHl5eXoPY2JlBQQGap/zMBgREVHNmWwAyszMxM6dO/Hiiy/K065fv47S0lJ4enrqjfX09ER2dnaFy1q4cCHUarX88PHxqbW6q0t3GIxnghEREdWcyQagdevWwdXVtUqHzCoyffp05Obmyo+MjIyaF2hgbIQmIiIyHBulC6gOIQTWrFmD4cOHw9bWVp7eoEEDWFtb48qVK3rjr1y5gkaNGlW4PJVKBZVKVWv1GgIDEBERkeGY5B6g/fv3IzU1FaNHj9abbmtri06dOmHPnj3yNI1Ggz179qBbt251XaZBMQAREREZjqJ7gAoKCpCamiq/Tk9PR3JyMtzc3ODr64vp06fj8uXL+Pzzz/Xmi42NRdeuXdGuXbsyy5w0aRKio6MREhKCLl26YOnSpSgsLMTIkSNrfXtqk+5q0BcvAkVFgL29svUQERGZMkUD0LFjx/Doo4/KrydNmgQAiI6ORlxcHLKysnDx4kW9eXJzc7FlyxYsW7as3GUOGTIE165dw6xZs5CdnY2OHTsiMTGxTGO0qXF3B9zcgJs3gZQUoEMHpSsiIiIyXUZzHSBjYmzXAdLp1g34+Wdg0yZg4EClqyEiIjIuFnMdIEvDPiAiIiLDYAAyIQxAREREhsEAZEIYgIiIiAyDAciE6M4E49WgiYiIaoYByIQ0b679efMmcOOGsrUQERGZMgYgE+LgAOhuU8bDYERERNXHAGRi2AdERERUcwxAJoYBiIiIqOYYgEwMG6GJiIhqjgHIxHAPEBERUc0xAJkYXQBKSQE0GmVrISIiMlUMQCbGzw+oVw+4cwe4dEnpaoiIiEwTA5CJsbEBAgK0z3kYjIiIqHoYgEwQG6GJiIhqhgHIBLERmoiIqGYYgEwQAxAREVHNMACZIAYgIiKimmEAMkG6AHT+PFBcrGgpREREJokByAR5egIuLtrrAP31l9LVEBERmR4GIBMkSTwMRkREVBMMQCaKAYiIiKj6GIBMFAMQERFR9TEAmSgGICIioupjADJRuqtBMwARERFVHQOQiQoM1P68cgXIzVW2FiIiIlPDAGSinJ0BLy/tc+4FIiIiqhoGIBPGPiAiIqLqYQAyYQxARERE1cMAZMLYCE1ERFQ9DEAmTLcH6OxZZesgIiIyNQxAJuz+Q2BCKFsLERGRKWEAMmFNmwLW1kBhIZCVpXQ1REREpoMByITZ2mpDEMA+ICIioqpgADJxPBOMiIio6hiATJzuTDA2QhMREVUeA5CJ4x4gIiKiqmMAMnEMQERERFXHAGTidAEoLQ24e1fZWoiIiEwFA5CJ8/YGHByAe/eA8+eVroaIiMg0MACZOCsrXhGaiIioqhiAzAD7gIiIiKpG0QB04MABREREwNvbG5IkISEh4aHzFBcXY8aMGfDz84NKpYK/vz/WrFmjN2bp0qVo2bIl7O3t4ePjg9dffx137typpa1QHgMQERFR1dgoufLCwkIEBQVh1KhRGDBgQKXmGTx4MK5cuYLY2Fg0b94cWVlZ0Gg08vsbNmzAtGnTsGbNGnTv3h3nzp3DiBEjIEkSlixZUluboigGICIioqpRNACFh4cjPDy80uMTExOxf/9+pKWlwc3NDQDg7++vN+ann35CaGgohg0bJr8/dOhQ/PLLLwar29gwABEREVWNSfUAbdu2DSEhIVi0aBEaN26MFi1aYPLkySgqKpLHdO/eHb/++iuOHDkCAEhLS8OOHTvQv39/pcqudboAdPkyUFCgbC1ERESmQNE9QFWVlpaGgwcPws7ODvHx8bh+/TrGjBmDGzduYO3atQCAYcOG4fr16/j3v/8NIQTu3buHV199Ff/9738rXG5xcTGKi4vl13l5ebW+LYZUvz7QsCFw7RqQkgIEBytdERERkXEzqT1AGo0GkiRh/fr16NKlC/r3748lS5Zg3bp18l6gffv2YcGCBfj4449x/PhxbN26Fdu3b8f8+fMrXO7ChQuhVqvlh4+PT11tksHwMBgREVHlmVQA8vLyQuPGjaFWq+VprVu3hhACly5dAgDMnDkTw4cPx4svvoj27dvjmWeewYIFC7Bw4UK9Zun7TZ8+Hbm5ufIjIyOjTrbHkBiAiIiIKs+kAlBoaCgyMzNRcF+jy7lz52BlZYUmTZoAAG7fvg0rK/3Nsra2BgAIIcpdrkqlgouLi97D1DAAERERVZ6iAaigoADJyclITk4GAKSnpyM5ORkXL14EoN0z88ILL8jjhw0bBnd3d4wcORKnT5/GgQMHMGXKFIwaNQr29vYAgIiICKxcuRIbN25Eeno6kpKSMHPmTERERMhByBy1bKn9yatBExERPZyiTdDHjh3Do48+Kr+eNGkSACA6OhpxcXHIysqSwxAAODk5ISkpCePGjUNISAjc3d0xePBgvPXWW/KYN998E5Ik4c0338Tly5fRsGFDRERE4O233667DVPA/XuAhAAkSdl6iIiIjJkkKjouZMHy8vKgVquRm5trMofD7tzR3hRVCODKFcDDQ+mKiIiI6lZV/n5X6xBYRkaG3HQMAEeOHMHEiROxevXq6iyODMDODvDz0z5nHxAREdGDVSsADRs2DHv37gUAZGdn44knnsCRI0cwY8YMzJs3z6AFUuWxEZqIiKhyqhWAfv/9d3Tp0gUA8L///Q/t2rXDTz/9hPXr1yMuLs6Q9VEV6BqhGYCIiIgerFoB6O7du1CpVACA3bt34z//+Q8AoFWrVsjKyjJcdVQluj1APBOMiIjowaoVgNq2bYtPPvkEP/74I5KSktCvXz8AQGZmJtzd3Q1aIFUeD4ERERFVTrUC0LvvvotVq1ahd+/eGDp0KIKCggBob1aqOzRGdU8XgFJTgdJSZWshIiIyZtU+Db60tBR5eXmoX7++PO38+fNwcHCAh4mfg22Kp8ED2tDj6AgUFwNpaUDTpkpXREREVHdq/TT4oqIiFBcXy+HnwoULWLp0Kc6ePWvy4ceUWVsDgYHa5zwMRkREVLFqBaCnn34an3/+OQAgJycHXbt2xfvvv4/IyEisXLnSoAVS1bARmoiI6OGqFYCOHz+OHj16AAA2b94MT09PXLhwAZ9//jmWL19u0AKpatgITURE9HDVCkC3b9+Gs7MzAOD777/HgAEDYGVlhX/961+4cOGCQQukqmEAIiIierhqBaDmzZsjISEBGRkZ2LVrF/r06QMAuHr1qkk1DZsjBiAiIqKHq1YAmjVrFiZPngx/f3906dIF3bp1A6DdGxQcHGzQAqlqdFeDvngRKCpSthYiIiJjVe3T4LOzs5GVlYWgoCBYWWlz1JEjR+Di4oJWrVoZtMi6ZqqnwQPau8G7uwO3bgG//Qa0b690RURERHWj1k+DB4BGjRohODgYmZmZ8p3hu3TpYvLhx9RJEg+DERERPUy1ApBGo8G8efOgVqvh5+cHPz8/uLq6Yv78+dBoNIaukaqIAYiIiOjBbKoz04wZMxAbG4t33nkHoaGhAICDBw9izpw5uHPnDt5++22DFklVwwBERET0YNUKQOvWrcNnn30m3wUeADp06IDGjRtjzJgxDEAKYwAiIiJ6sGodArt582a5vT6tWrXCzZs3a1wU1YzuTDBeDZqIiKh81QpAQUFB+PDDD8tM//DDD9GhQ4caF0U107y59ueNG9oHERER6avWIbBFixbhySefxO7du+VrAB0+fBgZGRnYsWOHQQukqnN0BJo0AS5dAlJStKfFExER0d+qtQeoV69eOHfuHJ555hnk5OQgJycHAwYMwB9//IEvvvjC0DVSNbAPiIiIqGLV2gMEAN7e3mWanU+ePInY2FisXr26xoVRzbRoAfzwAwMQERFReap9IUQybmyEJiIiqhgDkJniITAiIqKKMQCZKV0ASkkBeHFuIiIifVXqARowYMAD38/JyalJLWRA/v6AjY32jvCXLwM+PkpXREREZDyqFIDUavVD33/hhRdqVBAZho0NEBCg7QE6d44BiIiI6H5VCkBr166trTqoFrRsqQ1AZ88Cjz+udDVERETGgz1AZoyN0EREROVjADJjDEBERETlYwAyYwxARERE5WMAMmO6AJSeDpSUKFsLERGRMWEAMmONGgHOztrrAKWlKV0NERGR8WAAMmOS9PdeIN4Sg4iI6G8MQGaOfUBERERlMQCZOQYgIiKishiAzBwDEBERUVkMQGauZUvtTwYgIiKivzEAmbnAQO3P7GwgL0/ZWoiIiIwFA5CZc3HRng4PcC8QERGRDgOQBWAfEBERkT5FA9CBAwcQEREBb29vSJKEhISEh85TXFyMGTNmwM/PDyqVCv7+/lizZo3emJycHMTExMDLywsqlQotWrTAjh07amkrjB8DEBERkT4bJVdeWFiIoKAgjBo1CgMGDKjUPIMHD8aVK1cQGxuL5s2bIysrCxqNRn6/pKQETzzxBDw8PLB582Y0btwYFy5cgKuray1thfFjIzQREZE+RQNQeHg4wsPDKz0+MTER+/fvR1paGtzc3AAA/v7+emPWrFmDmzdv4qeffkK9evXKHWNpeDVoIiIifSbVA7Rt2zaEhIRg0aJFaNy4MVq0aIHJkyejqKhIb0y3bt0QExMDT09PtGvXDgsWLEBpaWmFyy0uLkZeXp7ew5zcfwhMCGVrISIiMgaK7gGqqrS0NBw8eBB2dnaIj4/H9evXMWbMGNy4cQNr166Vx/zwww+IiorCjh07kJqaijFjxuDu3buYPXt2uctduHAh5s6dW5ebUqeaNQOsrICCAu3p8F5eSldERESkLEkI49gnIEkS4uPjERkZWeGYPn364Mcff0R2djbUajUAYOvWrRg4cCAKCwthb2+PFi1a4M6dO0hPT4e1tTUAYMmSJXjvvfeQlZVV7nKLi4tRXFwsv87Ly4OPjw9yc3Ph4uJiuI1UUPPmwF9/Afv2Ab16KV0NERGR4eXl5UGtVlfq77dJHQLz8vJC48aN5fADAK1bt4YQApcuXZLHtGjRQg4/ujHZ2dkoKSkpd7kqlQouLi56D3PDRmgiIqK/mVQACg0NRWZmJgoKCuRp586dg5WVFZo0aSKPSU1N1Tsz7Ny5c/Dy8oKtrW2d12ws2AhNRET0N0UDUEFBAZKTk5GcnAwASE9PR3JyMi5evAgAmD59Ol544QV5/LBhw+Du7o6RI0fi9OnTOHDgAKZMmYJRo0bB3t4eAPDaa6/h5s2bmDBhAs6dO4ft27djwYIFiImJqfPtMya8FhAREdHfFA1Ax44dQ3BwMIKDgwEAkyZNQnBwMGbNmgUAyMrKksMQADg5OSEpKQk5OTkICQlBVFQUIiIisHz5cnmMj48Pdu3ahaNHj6JDhw4YP348JkyYgGnTptXtxhkZBiAiIqK/GU0TtDGpShOVqcjIAHx9ARsboKhI+5OIiMicmG0TNFVf48aAvT1w7x5w/rzS1RARESmLAchCWFmxEZqIiEiHAciCsA+IiIhIiwHIgjAAERERaTEAWRAGICIiIi0GIAvCAERERKTFAGRBdAHo0iWgsFDZWoiIiJTEAGRB3NyABg20z1NSlK2FiIhISQxAFoaHwYiIiBiALA4DEBEREQOQxWEAIiIiYgCyOC1ban/yatBERGTJGIAszP17gHgbXCIislQMQBYmIACQJCAnB7h+XelqiIiIlMEAZGHs7QFfX+1z9gEREZGlYgCyQGyEJiIiS8cAZIF0jdAMQEREZKkYgCyQbg8QzwQjIiJLxQBkgXgIjIiILB0DkAXSBaDUVKC0VNlaiIiIlMAAZIF8fQGVCiguBjIylK6GiIio7jEAWSBra6B5c+1zHgYjIiJLxABkodgITURElowByEKxEZqIiCwZA5CFYgAiIiJLxgBkoRiAiIjIkjEAWSjd1aAvXADu3FG2FiIiorrGAGShGjQAXF0BIbTXAyIiIrIkDEAWSpJ4GIyIiCwXA5AFYwAiIiJLxQBkwRiAiIjIUjEAWTAGICIislQMQBZMdyYYrwZNRESWhgHIgunuB3b9OnDzprK1EBER1SUGIAvm5AQ0bqx9npKibC1ERER1iQHIwrEPiIiILBEDkIVjACIiIkvEAGTh2AhNRESWiAHIwnEPEBERWSIGIAunC0ApKYBGo2wtREREdYUByML5+wM2NsDt20BmptLVEBER1Q0GIAtXrx7QrJn2OQ+DERGRpVA0AB04cAARERHw9vaGJElISEh46DzFxcWYMWMG/Pz8oFKp4O/vjzVr1pQ7duPGjZAkCZGRkYYt3MywEZqIiCyNjZIrLywsRFBQEEaNGoUBAwZUap7BgwfjypUriI2NRfPmzZGVlQVNOc0r58+fx+TJk9GjRw9Dl2122AhNRESWRtEAFB4ejvDw8EqPT0xMxP79+5GWlgY3NzcAgL+/f5lxpaWliIqKwty5c/Hjjz8iJyfHQBWbJwYgIiKyNCbVA7Rt2zaEhIRg0aJFaNy4MVq0aIHJkyejqKhIb9y8efPg4eGB0aNHV2q5xcXFyMvL03tYEgYgIiKyNIruAaqqtLQ0HDx4EHZ2doiPj8f169cxZswY3LhxA2vXrgUAHDx4ELGxsUhOTq70chcuXIi5c+fWUtXGTxeA0tOBkhLA1lbZeoiIiGqbSe0B0mg0kCQJ69evR5cuXdC/f38sWbIE69atQ1FREfLz8zF8+HB8+umnaNCgQaWXO336dOTm5sqPjIyMWtwK4+Plpb0xammpNgQRERGZO5PaA+Tl5YXGjRtDrVbL01q3bg0hBC5duoTCwkKcP38eERER8vu6BmkbGxucPXsWAQEBZZarUqmgUqlqfwOMlCRp9wIdP649E0x3VhgREZG5MqkAFBoaik2bNqGgoABOTk4AgHPnzsHKygpNmjSBJEk4deqU3jxvvvkm8vPzsWzZMvj4+ChRtknQBSD2ARERkSVQNAAVFBQgNTVVfp2eno7k5GS4ubnB19cX06dPx+XLl/H5558DAIYNG4b58+dj5MiRmDt3Lq5fv44pU6Zg1KhRsLe3BwC0a9dObx2urq7lTid9bIQmIiJLomgP0LFjxxAcHIzg4GAAwKRJkxAcHIxZs2YBALKysnDx4kV5vJOTE5KSkpCTk4OQkBBERUUhIiICy5cvV6R+c8IARERElkQSQgilizA2eXl5UKvVyM3NhYuLi9Ll1Iljx4DOnbUN0bwnGBERmaKq/P02qbPAqPYEBmp/ZmUBFnYZJCIiskAMQAQAUKsBT0/t85QUZWshIiKqbQxAJGMfEBERWQoGIJIxABERkaVgACKZ7gKIDEBERGTuGIBIptsDdPassnUQERHVNgYgkt1/CIwXRyAiInPGAESyZs0AKysgPx+4ckXpaoiIiGoPAxDJVCrA31/7nH1ARERkzhiASA8boYmIyBIwAJEeNkITEZElYAAiPbwWEBERWQIGINLDAERERJaAAYj06ALQX38B9+4pWwsREVFtYQAiPU2aAPb2wN27wIULSldDRERUOxiASI+VFRAYqH3ORmgiIjJXDEBUBvuAiIjI3DEAURkMQEREZO4YgKgMBiAiIjJ3DEBUBgMQERGZOwYgKkN3O4yMDKCwUNlaiIiIagMDEJXh5ga4u2ufp6YqWwsREVFtYACicvEwGBERmTMGICoXAxAREZkzBiAqFwMQERGZMwYgKpeuEZoBiIiIzBEDEJVLtwfo7FlACGVrISIiMjQGICpX8+ban7duATduKFsLERGRoTEAUbns7QFfX+1zHgYjIiJzwwBEFWIjNBERmSsGIKoQG6GJiMhcMQBRhe5vhCYiIjInDEBUIR4CIyIic8UARBXSBaCUFECjUbYWIiIiQ2IAogr5+QG2tkBxsfbO8EREROaCAYgqZG399/WAeBiMiIjMCQMQPRAboYmIyBwxANEDsRGaiIjMEQMQPRADEBERmSMGIHogBiAiIjJHDED0QLqrQZ8/rz0bjIiIyBwwANEDNWwIqNWAEEBqqtLVEBERGYaiAejAgQOIiIiAt7c3JElCQkLCQ+cpLi7GjBkz4OfnB5VKBX9/f6xZs0Z+/9NPP0WPHj1Qv3591K9fH2FhYThy5EgtboV5kyQeBiMiIvOjaAAqLCxEUFAQPvroo0rPM3jwYOzZswexsbE4e/YsvvrqK7TUHacBsG/fPgwdOhR79+7F4cOH4ePjgz59+uDy5cu1sQkWgQGIiIjMjY2SKw8PD0d4eHilxycmJmL//v1IS0uDm5sbAMDf319vzPr16/Vef/bZZ9iyZQv27NmDF154ocY1WyIGICIiMjcm1QO0bds2hISEYNGiRWjcuDFatGiByZMno6ioqMJ5bt++jbt378qBqTzFxcXIy8vTe9DfGICIiMjcKLoHqKrS0tJw8OBB2NnZIT4+HtevX8eYMWNw48YNrF27ttx5pk6dCm9vb4SFhVW43IULF2Lu3Lm1VbbJ0x1h5NWgiYjIXJjUHiCNRgNJkrB+/Xp06dIF/fv3x5IlS7Bu3bpy9wK988472LhxI+Lj42FnZ1fhcqdPn47c3Fz5kcE7f+oJDNT+vHYNuHVL2VqIiIgMwaQCkJeXFxo3bgy1Wi1Pa926NYQQuHTpkt7YxYsX45133sH333+PDh06PHC5KpUKLi4ueg/6m5MT4O2tfZ6SomwtREREhmBSASg0NBSZmZkoKCiQp507dw5WVlZo0qSJPG3RokWYP38+EhMTERISokSpZod9QEREZE4UDUAFBQVITk5GcnIyACA9PR3Jycm4ePEiAO2hqfvP3Bo2bBjc3d0xcuRInD59GgcOHMCUKVMwatQo2NvbAwDeffddzJw5E2vWrIG/vz+ys7ORnZ2tF5qo6hiAiIjInCgagI4dO4bg4GAEBwcDACZNmoTg4GDMmjULAJCVlSWHIQBwcnJCUlIScnJyEBISgqioKERERGD58uXymJUrV6KkpAQDBw6El5eX/Fi8eHHdbpyZYSM0ERGZE0kIIZQuwtjk5eVBrVYjNzeX/UD/33ffARERQMeOwIkTSldDRERUVlX+fptUDxAp5/5DYIzMRERk6hiAqFKaNgWsrYHbt4HMTKWrISIiqhkGIKqUevWAZs20z9kITUREpo4BiCpN1wjNAERERKaOAYgqTdcHxDPBiIjI1DEAUaXxWkBERGQuGICo0hiAiIjIXDAAUaXpAlBaGnD3rrK1EBER1QQDEFWatzfg6AiUlgLp6UpXQ0REVH0MQFRpksRGaCIiMg8MQFQl7AMiIiJzwABEVcIARERE5oABiKqEAYiIiMwBAxBVCa8GTURE5oABiKokMFD7MzMTyM9XthYiIqLqYgCiKnF1BTw8tM9TUhQthYiIqNoYgKjK2AdERESmjgGIqowBiIiITB0DEFUZG6GJiMjU2ShdgEW5cwfIyqqbdUlSrS26oyvgByDzsITEVZVbdV1MQ3njyhsm/XOMKGeUYdXFOipcdyW/Ckp9buVNE5AgyvmV3T9N97w60+pief/0oM/h/vfKfD8reK+y4yq7rgctzxyU9xlW5r3KjKvs8h407kHTK1NfVb6PVZ1uiGWUN13tocKjw7wqv1IDYwCqS8nJQLduSldRY30AnAeANACvKloKERGZqFNO3YBhPym2fgagumRlBTg41P56qhLjq7N4AHdLROVWU8lSlNs/olt/Df+5a4B/Lte4hrpQy98toO72lpX72zbgRyBB1N1nWtG/uutm7WbtgZ+gCfwna8zsXFWKrp8BqC516QIUFipdRY1JAGyVLoKIiExaoMLrZxM0ERERWRwGICIiIrI4DEBERERkcRiAiIiIyOIwABEREZHFYQAiIiIii8MARERERBaHAYiIiIgsDgMQERERWRwGICIiIrI4DEBERERkcRiAiIiIyOIwABEREZHFYQAiIiIii2OjdAHGSAgBAMjLy1O4EiIiIqos3d9t3d/xB2EAKkd+fj4AwMfHR+FKiIiIqKry8/OhVqsfOEYSlYlJFkaj0SAzMxPOzs6QJEnpcoxSXl4efHx8kJGRARcXF6XLsXj8PIwLPw/jw8/EuNTW5yGEQH5+Pry9vWFl9eAuH+4BKoeVlRWaNGmidBkmwcXFhf8zMSL8PIwLPw/jw8/EuNTG5/GwPT86bIImIiIii8MARERERBaHAYiqRaVSYfbs2VCpVEqXQuDnYWz4eRgffibGxRg+DzZBExERkcXhHiAiIiKyOAxAREREZHEYgIiIiMjiMAARERGRxWEAoipZuHAhOnfuDGdnZ3h4eCAyMhJnz55VuiwC8M4770CSJEycOFHpUiza5cuX8fzzz8Pd3R329vZo3749jh07pnRZFqm0tBQzZ85E06ZNYW9vj4CAAMyfP79S94mimjtw4AAiIiLg7e0NSZKQkJCg974QArNmzYKXlxfs7e0RFhaGlJSUOquPAYiqZP/+/YiJicHPP/+MpKQk3L17F3369EFhYaHSpVm0o0ePYtWqVejQoYPSpVi0W7duITQ0FPXq1cPOnTtx+vRpvP/++6hfv77SpVmkd999FytXrsSHH36IM2fO4N1338WiRYuwYsUKpUuzCIWFhQgKCsJHH31U7vuLFi3C8uXL8cknn+CXX36Bo6Mj+vbtizt37tRJfTwNnmrk2rVr8PDwwP79+9GzZ0+ly7FIBQUFeOSRR/Dxxx/jrbfeQseOHbF06VKly7JI06ZNw6FDh/Djjz8qXQoBeOqpp+Dp6YnY2Fh52rPPPgt7e3t8+eWXClZmeSRJQnx8PCIjIwFo9/54e3vjjTfewOTJkwEAubm58PT0RFxcHJ577rlar4l7gKhGcnNzAQBubm4KV2K5YmJi8OSTTyIsLEzpUizetm3bEBISgkGDBsHDwwPBwcH49NNPlS7LYnXv3h179uzBuXPnAAAnT57EwYMHER4ernBllJ6ejuzsbL3/b6nVanTt2hWHDx+ukxp4M1SqNo1Gg4kTJyI0NBTt2rVTuhyLtHHjRhw/fhxHjx5VuhQCkJaWhpUrV2LSpEn473//i6NHj2L8+PGwtbVFdHS00uVZnGnTpiEvLw+tWrWCtbU1SktL8fbbbyMqKkrp0ixednY2AMDT01Nvuqenp/xebWMAomqLiYnB77//joMHDypdikXKyMjAhAkTkJSUBDs7O6XLIWj/URASEoIFCxYAAIKDg/H777/jk08+YQBSwP/+9z+sX78eGzZsQNu2bZGcnIyJEyfC29ubnwfxEBhVz9ixY/Hdd99h7969aNKkidLlWKRff/0VV69exSOPPAIbGxvY2Nhg//79WL58OWxsbFBaWqp0iRbHy8sLbdq00ZvWunVrXLx4UaGKLNuUKVMwbdo0PPfcc2jfvj2GDx+O119/HQsXLlS6NIvXqFEjAMCVK1f0pl+5ckV+r7YxAFGVCCEwduxYxMfH44cffkDTpk2VLsliPf744zh16hSSk5PlR0hICKKiopCcnAxra2ulS7Q4oaGhZS4Lce7cOfj5+SlUkWW7ffs2rKz0/8xZW1tDo9EoVBHpNG3aFI0aNcKePXvkaXl5efjll1/QrVu3OqmBh8CoSmJiYrBhwwZ88803cHZ2lo/VqtVq2NvbK1ydZXF2di7Te+Xo6Ah3d3f2ZCnk9ddfR/fu3bFgwQIMHjwYR44cwerVq7F69WqlS7NIERERePvtt+Hr64u2bdvixIkTWLJkCUaNGqV0aRahoKAAqamp8uv09HQkJyfDzc0Nvr6+mDhxIt566y0EBgaiadOmmDlzJry9veUzxWqdIKoCAOU+1q5dq3RpJITo1auXmDBhgtJlWLRvv/1WtGvXTqhUKtGqVSuxevVqpUuyWHl5eWLChAnC19dX2NnZiWbNmokZM2aI4uJipUuzCHv37i3370V0dLQQQgiNRiNmzpwpPD09hUqlEo8//rg4e/ZsndXH6wARERGRxWEPEBEREVkcBiAiIiKyOAxAREREZHEYgIiIiMjiMAARERGRxWEAIiIiIovDAEREREQWhwGIiKgCkiQhISFB6TKIqBYwABGRURoxYgQkSSrz6Nevn9KlEZEZ4L3AiMho9evXD2vXrtWbplKpFKqGiMwJ9wARkdFSqVRo1KiR3qN+/foAtIenVq5cifDwcNjb26NZs2bYvHmz3vynTp3CY489Bnt7e7i7u+Pll19GQUGB3pg1a9agbdu2UKlU8PLywtixY/Xev379Op555hk4ODggMDAQ27Ztk9+7desWoqKi0LBhQ9jb2yMwMLBMYCMi48QAREQma+bMmXj22Wdx8uRJREVF4bnnnsOZM2cAAIWFhejbty/q16+Po0ePYtOmTdi9e7dewFm5ciViYmLw8ssv49SpU9i2bRuaN2+ut465c+di8ODB+O2339C/f39ERUXh5s2b8vpPnz6NnTt34syZM1i5ciUaNGhQd78AIqq+OrvtKhFRFURHRwtra2vh6Oio93j77beFEEIAEK+++qrePF27dhWvvfaaEEKI1atXi/r164uCggL5/e3btwsrKyuRnZ0thBDC29tbzJgxo8IaAIg333xTfl1QUCAAiJ07dwohhIiIiBAjR440zAYTUZ1iDxARGa1HH30UK1eu1Jvm5uYmP+/WrZvee926dUNycjIA4MyZMwgKCoKjo6P8fmhoKDQaDc6ePQtJkpCZmYnHH3/8gTV06NBBfu7o6AgXFxdcvXoVAPDaa6/h2WefxfHjx9GnTx9ERkaie/fu1dpWIqpbDEBEZLQcHR3LHJIyFHt7+0qNq1evnt5rSZKg0WgAAOHh4bhw4QJ27NiBpKQkPP7444iJicHixYsNXi8RGRZ7gIjIZP38889lXrdu3RoA0Lp1a5w8eRKFhYXy+4cOHYKVlRVatmwJZ2dn+Pv7Y8+ePTWqoWHDhoiOjsaXX36JpUuXYvXq1TVaHhHVDe4BIiKjVVxcjOzsbL1pNjY2cqPxpk2bEBISgn//+99Yv349jhw5gtjYWABAVFQUZs+ejejoaMyZMwfXrl3DuHHjMHz4cHh6egIA5syZg1dffRUeHh4IDw9Hfn4+Dh06hHHjxlWqvlmzZqFTp05o27YtiouL8d1338kBjIiMGwMQERmtxMREeHl56U1r2bIl/vzzTwDaM7Q2btyIMWPGwMvLC1999RXatGkDAHBwcMCuXbswYcIEdO7cGQ4ODnj22WexZMkSeVnR0dG4c+cOPvjgA0yePBkNGjTAwIEDK12fra0tpk+fjvPnz8Pe3h49evTAxo0bDbDlRFTbJCGEULoIIqKqkiQJ8fHxiIyMVLoUIjJB7AEiIiIii8MARERERBaHPUBEZJJ49J6IaoJ7gIiIiMjiMAARERGRxWEAIiIiIovDAEREREQWhwGIiIiILA4DEBEREVkcBiAiIiKyOAxAREREZHEYgIiIiMji/D+A05v1MfWTAQAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Accessing the loss values\n",
        "train_loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "# Plotting the loss values\n",
        "epochs = range(1, len(train_loss) + 1)\n",
        "\n",
        "plt.plot(epochs, train_loss, 'b', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'r', label='Validation loss')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "M6Sr9EbTnBGM",
        "outputId": "abefd599-98cd-4780-94be-c42479d8ce40"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 5000 images belonging to 5 classes.\n",
            "10/10 [==============================] - 7s 602ms/step - loss: 1.6094 - accuracy: 0.2000\n",
            "Test loss: 1.6094413995742798\n",
            "Test accuracy: 0.20000000298023224\n"
          ]
        }
      ],
      "source": [
        "score = model.evaluate(test_generator)\n",
        "print('Test loss:', score[0])\n",
        "print('Test accuracy:', score[1])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vZB64EuwEeaH"
      },
      "source": [
        "# Over Fitting Model\n",
        "To induce overfitting in our model, we can increase its capacity or complexity. Here are a few modifications we can make to our model to encourage overfitting:\n",
        "\n",
        "1. Increase the number of filters in the convolutional layers: By increasing the number of filters, the model can learn more complex features from the input data.\n",
        "\n",
        "2. Add more convolutional layers: Increasing the depth of the model allows it to learn more intricate patterns in the data.\n",
        "\n",
        "3. Increase the number of neurons in the dense layer: By adding more neurons, the model can capture more complex relationships in the data."
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "train_datagen = ImageDataGenerator(rescale=1./255)\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    '/content/sample_data/miotcd/HW02-Practical/Dataset/train',\n",
        "    target_size=(128, 128),\n",
        "    batch_size=128,\n",
        "    class_mode='categorical'\n",
        ")\n",
        "\n",
        "validation_datagen = ImageDataGenerator(rescale=1./255)\n",
        "validation_generator = validation_datagen.flow_from_directory(\n",
        "    '/content/sample_data/miotcd/HW02-Practical/Dataset/valid',\n",
        "    target_size=(128, 128),\n",
        "    batch_size=128,\n",
        "    class_mode='categorical'\n",
        ")"
      ],
      "metadata": {
        "id": "K0NtL3Plu4z7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NtTxINwaErcU"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Conv2D(64, (3, 3),activation='relu', input_shape=(128, 128, 3)))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "model.add(Conv2D(128, (3, 3),activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "model.add(Conv2D(256, (3, 3),activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(512, activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(256, activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(5, activation='softmax'))\n",
        "\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QP4MOzydE6_A",
        "outputId": "af8ce808-2d7d-49b3-a38f-09d206cc791f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "117/117 [==============================] - 71s 467ms/step - loss: 1.1195 - accuracy: 0.5616 - val_loss: 0.8049 - val_accuracy: 0.7262\n",
            "Epoch 2/10\n",
            "117/117 [==============================] - 53s 454ms/step - loss: 0.6125 - accuracy: 0.7805 - val_loss: 0.5589 - val_accuracy: 0.8039\n",
            "Epoch 3/10\n",
            "117/117 [==============================] - 61s 519ms/step - loss: 0.4247 - accuracy: 0.8536 - val_loss: 0.4877 - val_accuracy: 0.8317\n",
            "Epoch 4/10\n",
            "117/117 [==============================] - 53s 450ms/step - loss: 0.3164 - accuracy: 0.8934 - val_loss: 0.4300 - val_accuracy: 0.8558\n",
            "Epoch 5/10\n",
            "117/117 [==============================] - 61s 518ms/step - loss: 0.2245 - accuracy: 0.9248 - val_loss: 0.4789 - val_accuracy: 0.8580\n",
            "Epoch 6/10\n",
            "117/117 [==============================] - 60s 515ms/step - loss: 0.1573 - accuracy: 0.9477 - val_loss: 0.4165 - val_accuracy: 0.8812\n",
            "Epoch 7/10\n",
            "117/117 [==============================] - 53s 453ms/step - loss: 0.1264 - accuracy: 0.9582 - val_loss: 0.4351 - val_accuracy: 0.8766\n",
            "Epoch 8/10\n",
            "117/117 [==============================] - 61s 522ms/step - loss: 0.1148 - accuracy: 0.9631 - val_loss: 0.4935 - val_accuracy: 0.8722\n",
            "Epoch 9/10\n",
            "117/117 [==============================] - 62s 527ms/step - loss: 0.0852 - accuracy: 0.9734 - val_loss: 0.5273 - val_accuracy: 0.8720\n",
            "Epoch 10/10\n",
            "117/117 [==============================] - 61s 523ms/step - loss: 0.0778 - accuracy: 0.9724 - val_loss: 0.5061 - val_accuracy: 0.8720\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(\n",
        "    train_generator,\n",
        "    steps_per_epoch=train_generator.n // train_generator.batch_size,\n",
        "    epochs=10,\n",
        "    validation_data=validation_generator,\n",
        "    validation_steps=validation_generator.n // validation_generator.batch_size\n",
        "    )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "D4LI3B4jKwTS",
        "outputId": "3c7c289d-3f8a-4da6-9979-3e6eb4551dac"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABmP0lEQVR4nO3dd3zM9x8H8NclkSUDQQYRxIoVxCipHbUaDVWjSqhVe7aoPUqtUntURYsatVeIFEXVjlGxaqWILZGIhLvv74/P7y452cldvjdez8fjHrn73ve+3/cl4V75fD9DIUmSBCIiIiITYSF3AURERES6xHBDREREJoXhhoiIiEwKww0RERGZFIYbIiIiMikMN0RERGRSGG6IiIjIpDDcEBERkUlhuCEiIiKTwnBDlIe6d++OkiVL5ui1kyZNgkKh0G1BBubOnTtQKBQICQnJ83MrFApMmjRJ8zgkJAQKhQJ37tzJ9LUlS5ZE9+7ddVpPbn5XiMwdww0RxAdbVm6HDx+Wu1SzN3jwYCgUCty8eTPdfcaOHQuFQoGLFy/mYWXZ9+DBA0yaNAkRERFyl6KhDphz5syRuxSiHLOSuwAiQ/Drr79qPf7ll18QFhaWaruPj0+uzrNy5UqoVKocvXbcuHEYPXp0rs5vCrp06YKFCxdi/fr1mDBhQpr7/Pbbb6hSpQqqVq2a4/N07doVnTp1go2NTY6PkZkHDx5g8uTJKFmyJKpVq6b1XG5+V4jMHcMNEYAvvvhC6/Hff/+NsLCwVNvf9/r1a9jb22f5PPny5ctRfQBgZWUFKyv+k61Tpw7KlCmD3377Lc1wc+LECdy+fRvff/99rs5jaWkJS0vLXB0jN3Lzu0Jk7nhZiiiLGjVqhMqVK+Ps2bNo0KAB7O3t8e233wIAduzYgdatW8PDwwM2Njbw9vbG1KlToVQqtY7xfj+KlJcAVqxYAW9vb9jY2KBWrVo4ffq01mvT6nOjUCgwcOBAbN++HZUrV4aNjQ0qVaqE0NDQVPUfPnwYNWvWhK2tLby9vbF8+fIs9+M5evQoPvvsM5QoUQI2Njbw9PTEsGHDkJCQkOr9OTg44P79+wgKCoKDgwOKFCmCkSNHpvpevHz5Et27d4ezszMKFCiA4OBgvHz5MtNaANF6c/XqVZw7dy7Vc+vXr4dCoUDnzp2RlJSECRMmwM/PD87OzsifPz/q16+PQ4cOZXqOtPrcSJKEadOmoXjx4rC3t0fjxo3xzz//pHrt8+fPMXLkSFSpUgUODg5wcnJCy5YtceHCBc0+hw8fRq1atQAAPXr00Fz6VPc3SqvPTXx8PEaMGAFPT0/Y2NigfPnymDNnDiRJ0tovO78XOfX48WP07NkTrq6usLW1ha+vL9asWZNqvw0bNsDPzw+Ojo5wcnJClSpV8OOPP2qef/v2LSZPnoyyZcvC1tYWLi4u+PDDDxEWFqazWsn88M9Aomx49uwZWrZsiU6dOuGLL76Aq6srAPFB6ODggOHDh8PBwQF//PEHJkyYgNjYWMyePTvT465fvx6vXr1C3759oVAoMGvWLLRr1w63bt3K9C/4Y8eOYevWrejfvz8cHR2xYMECfPrpp7h37x5cXFwAAOfPn0eLFi3g7u6OyZMnQ6lUYsqUKShSpEiW3vfmzZvx+vVr9OvXDy4uLjh16hQWLlyI//77D5s3b9baV6lUonnz5qhTpw7mzJmDgwcPYu7cufD29ka/fv0AiJDwySef4NixY/jqq6/g4+ODbdu2ITg4OEv1dOnSBZMnT8b69etRo0YNrXNv2rQJ9evXR4kSJfD06VP89NNP6Ny5M3r37o1Xr15h1apVaN68OU6dOpXqUlBmJkyYgGnTpqFVq1Zo1aoVzp07h48++ghJSUla+926dQvbt2/HZ599hlKlSuHRo0dYvnw5GjZsiCtXrsDDwwM+Pj6YMmUKJkyYgD59+qB+/foAgHr16qV5bkmS0KZNGxw6dAg9e/ZEtWrVsH//fnz99de4f/8+5s2bp7V/Vn4vciohIQGNGjXCzZs3MXDgQJQqVQqbN29G9+7d8fLlSwwZMgQAEBYWhs6dO6Np06aYOXMmACAyMhLHjx/X7DNp0iTMmDEDvXr1Qu3atREbG4szZ87g3LlzaNasWa7qJDMmEVEqAwYMkN7/59GwYUMJgLRs2bJU+79+/TrVtr59+0r29vbSmzdvNNuCg4MlLy8vzePbt29LACQXFxfp+fPnmu07duyQAEi7du3SbJs4cWKqmgBI1tbW0s2bNzXbLly4IAGQFi5cqNkWGBgo2dvbS/fv39dsu3HjhmRlZZXqmGlJ6/3NmDFDUigU0t27d7XeHwBpypQpWvtWr15d8vPz0zzevn27BECaNWuWZtu7d++k+vXrSwCk1atXZ1pTrVq1pOLFi0tKpVKzLTQ0VAIgLV++XHPMxMRErde9ePFCcnV1lb788kut7QCkiRMnah6vXr1aAiDdvn1bkiRJevz4sWRtbS21bt1aUqlUmv2+/fZbCYAUHBys2fbmzRutuiRJ/KxtbGy0vjenT59O9/2+/7ui/p5NmzZNa7/27dtLCoVC63cgq78XaVH/Ts6ePTvdfebPny8BkNauXavZlpSUJNWtW1dycHCQYmNjJUmSpCFDhkhOTk7Su3fv0j2Wr6+v1Lp16wxrIsouXpYiygYbGxv06NEj1XY7OzvN/VevXuHp06eoX78+Xr9+jatXr2Z63I4dO6JgwYKax+q/4m/dupXpawMCAuDt7a15XLVqVTg5OWleq1QqcfDgQQQFBcHDw0OzX5kyZdCyZctMjw9ov7/4+Hg8ffoU9erVgyRJOH/+fKr9v/rqK63H9evX13ove/fuhZWVlaYlBxB9XAYNGpSlegDRT+q///7Dn3/+qdm2fv16WFtb47PPPtMc09raGgCgUqnw/PlzvHv3DjVr1kzzklZGDh48iKSkJAwaNEjrUt7QoUNT7WtjYwMLC/Hfq1KpxLNnz+Dg4IDy5ctn+7xqe/fuhaWlJQYPHqy1fcSIEZAkCfv27dPantnvRW7s3bsXbm5u6Ny5s2Zbvnz5MHjwYMTFxeHIkSMAgAIFCiA+Pj7DS0wFChTAP//8gxs3buS6LiI1hhuibChWrJjmwzKlf/75B23btoWzszOcnJxQpEgRTWfkmJiYTI9bokQJrcfqoPPixYtsv1b9evVrHz9+jISEBJQpUybVfmltS8u9e/fQvXt3FCpUSNOPpmHDhgBSvz9bW9tUl7tS1gMAd+/ehbu7OxwcHLT2K1++fJbqAYBOnTrB0tIS69evBwC8efMG27ZtQ8uWLbWC4po1a1C1alVNf44iRYpgz549Wfq5pHT37l0AQNmyZbW2FylSROt8gAhS8+bNQ9myZWFjY4PChQujSJEiuHjxYrbPm/L8Hh4ecHR01NquHsGnrk8ts9+L3Lh79y7Kli2rCXDp1dK/f3+UK1cOLVu2RPHixfHll1+m6vczZcoUvHz5EuXKlUOVKlXw9ddfG/wQfjJ8DDdE2ZCyBUPt5cuXaNiwIS5cuIApU6Zg165dCAsL0/QxyMpw3vRG5UjvdRTV9WuzQqlUolmzZtizZw9GjRqF7du3IywsTNPx9f33l1cjjIoWLYpmzZphy5YtePv2LXbt2oVXr16hS5cumn3Wrl2L7t27w9vbG6tWrUJoaCjCwsLQpEkTvQ6znj59OoYPH44GDRpg7dq12L9/P8LCwlCpUqU8G96t79+LrChatCgiIiKwc+dOTX+hli1bavWtatCgAf7991/8/PPPqFy5Mn766SfUqFEDP/30U57VSaaHHYqJcunw4cN49uwZtm7digYNGmi23759W8aqkhUtWhS2trZpTnqX0UR4apcuXcL169exZs0adOvWTbM9N6NZvLy8EB4ejri4OK3Wm2vXrmXrOF26dEFoaCj27duH9evXw8nJCYGBgZrnf//9d5QuXRpbt27VupQ0ceLEHNUMADdu3EDp0qU12588eZKqNeT3339H48aNsWrVKq3tL1++ROHChTWPszPjtJeXFw4ePIhXr15ptd6oL3uq68sLXl5euHjxIlQqlVbrTVq1WFtbIzAwEIGBgVCpVOjfvz+WL1+O8ePHa1oOCxUqhB49eqBHjx6Ii4tDgwYNMGnSJPTq1SvP3hOZFrbcEOWS+i/klH8RJyUlYcmSJXKVpMXS0hIBAQHYvn07Hjx4oNl+8+bNVP000ns9oP3+JEnSGs6bXa1atcK7d++wdOlSzTalUomFCxdm6zhBQUGwt7fHkiVLsG/fPrRr1w62trYZ1n7y5EmcOHEi2zUHBAQgX758WLhwodbx5s+fn2pfS0vLVC0kmzdvxv3797W25c+fHwCyNAS+VatWUCqVWLRokdb2efPmQaFQZLn/lC60atUK0dHR2Lhxo2bbu3fvsHDhQjg4OGguWT579kzrdRYWFpqJFRMTE9Pcx8HBAWXKlNE8T5QTbLkhyqV69eqhYMGCCA4O1iwN8Ouvv+Zp839mJk2ahAMHDsDf3x/9+vXTfEhWrlw506n/K1SoAG9vb4wcORL379+Hk5MTtmzZkqu+G4GBgfD398fo0aNx584dVKxYEVu3bs12fxQHBwcEBQVp+t2kvCQFAB9//DG2bt2Ktm3bonXr1rh9+zaWLVuGihUrIi4uLlvnUs/XM2PGDHz88cdo1aoVzp8/j3379mm1xqjPO2XKFPTo0QP16tXDpUuXsG7dOq0WHwDw9vZGgQIFsGzZMjg6OiJ//vyoU6cOSpUqler8gYGBaNy4McaOHYs7d+7A19cXBw4cwI4dOzB06FCtzsO6EB4ejjdv3qTaHhQUhD59+mD58uXo3r07zp49i5IlS+L333/H8ePHMX/+fE3LUq9evfD8+XM0adIExYsXx927d7Fw4UJUq1ZN0z+nYsWKaNSoEfz8/FCoUCGcOXMGv//+OwYOHKjT90NmRp5BWkSGLb2h4JUqVUpz/+PHj0sffPCBZGdnJ3l4eEjffPONtH//fgmAdOjQIc1+6Q0FT2vYLd4bmpzeUPABAwakeq2Xl5fW0GRJkqTw8HCpevXqkrW1teTt7S399NNP0ogRIyRbW9t0vgvJrly5IgUEBEgODg5S4cKFpd69e2uGFqccxhwcHCzlz58/1evTqv3Zs2dS165dJScnJ8nZ2Vnq2rWrdP78+SwPBVfbs2ePBEByd3dPNfxapVJJ06dPl7y8vCQbGxupevXq0u7du1P9HCQp86HgkiRJSqVSmjx5suTu7i7Z2dlJjRo1ki5fvpzq+/3mzRtpxIgRmv38/f2lEydOSA0bNpQaNmyodd4dO3ZIFStW1AzLV7/3tGp89eqVNGzYMMnDw0PKly+fVLZsWWn27NlaQ9PV7yWrvxfvU/9Opnf79ddfJUmSpEePHkk9evSQChcuLFlbW0tVqlRJ9XP7/fffpY8++kgqWrSoZG1tLZUoUULq27ev9PDhQ80+06ZNk2rXri0VKFBAsrOzkypUqCB99913UlJSUoZ1EmVEIUkG9OclEeWpoKAgDsMlIpPDPjdEZuL9pRJu3LiBvXv3olGjRvIURESkJ2y5ITIT7u7u6N69O0qXLo27d+9i6dKlSExMxPnz51PN3UJEZMzYoZjITLRo0QK//fYboqOjYWNjg7p162L69OkMNkRkcthyQ0RERCaFfW6IiIjIpDDcEBERkUkxuz43KpUKDx48gKOjY7amPiciIiL5SJKEV69ewcPDI9Wire8zu3Dz4MEDeHp6yl0GERER5UBUVBSKFy+e4T5mF27U04JHRUXByclJ5mqIiIgoK2JjY+Hp6am1cGx6zC7cqC9FOTk5MdwQEREZmax0KWGHYiIiIjIpDDdERERkUhhuiIiIyKSYXZ8bIiLSLaVSibdv38pdBpkAa2vrTId5ZwXDDRER5YgkSYiOjsbLly/lLoVMhIWFBUqVKgVra+tcHYfhhoiIckQdbIoWLQp7e3tOjEq5op5k9+HDhyhRokSufp8YboiIKNuUSqUm2Li4uMhdDpmIIkWK4MGDB3j37h3y5cuX4+OwQzEREWWbuo+Nvb29zJWQKVFfjlIqlbk6DsMNERHlGC9FkS7p6veJ4YaIiIhMCsMNERFRLpUsWRLz58/P8v6HDx+GQqHQ+0izkJAQFChQQK/nMEQMN0REZDYUCkWGt0mTJuXouKdPn0afPn2yvH+9evXw8OFDODs75+h8lDGOltKhJ0+Ahw+BqlXlroSIiNLy8OFDzf2NGzdiwoQJuHbtmmabg4OD5r4kSVAqlbCyyvyjskiRItmqw9raGm5ubtl6DWUdW250ZPt2wNUV6N1b7kqIiCg9bm5umpuzszMUCoXm8dWrV+Ho6Ih9+/bBz88PNjY2OHbsGP7991988skncHV1hYODA2rVqoWDBw9qHff9y1IKhQI//fQT2rZtC3t7e5QtWxY7d+7UPP/+ZSn15aP9+/fDx8cHDg4OaNGihVYYe/fuHQYPHowCBQrAxcUFo0aNQnBwMIKCgrL1PVi6dCm8vb1hbW2N8uXL49dff9U8J0kSJk2ahBIlSsDGxgYeHh4YPHiw5vklS5agbNmysLW1haurK9q3b5+tc+cVhhsdqVMHkCTg1CnRekNEZG4kCYiPl+cmSbp7H6NHj8b333+PyMhIVK1aFXFxcWjVqhXCw8Nx/vx5tGjRAoGBgbh3716Gx5k8eTI6dOiAixcvolWrVujSpQueP3+e7v6vX7/GnDlz8Ouvv+LPP//EvXv3MHLkSM3zM2fOxLp167B69WocP34csbGx2L59e7be27Zt2zBkyBCMGDECly9fRt++fdGjRw8cOnQIALBlyxbMmzcPy5cvx40bN7B9+3ZUqVIFAHDmzBkMHjwYU6ZMwbVr1xAaGooGDRpk6/x5RjIzMTExEgApJiZG58euXVuSAElasULnhyYiMigJCQnSlStXpISEBM22uDjxf6Act7i47L+H1atXS87OzprHhw4dkgBI27dvz/S1lSpVkhYuXKh57OXlJc2bN0/zGIA0bty4FN+bOAmAtG/fPq1zvXjxQlMLAOnmzZua1yxevFhydXXVPHZ1dZVmz56tefzu3TupRIkS0ieffJLl91ivXj2pd+/eWvt89tlnUqtWrSRJkqS5c+dK5cqVk5KSklIda8uWLZKTk5MUGxub7vlyK63fK7XsfH6z5UaHPvlEfN2xQ946iIgo52rWrKn1OC4uDiNHjoSPjw8KFCgABwcHREZGZtpyUzVFB8z8+fPDyckJjx8/Tnd/e3t7eHt7ax67u7tr9o+JicGjR49Qu3ZtzfOWlpbw8/PL1nuLjIyEv7+/1jZ/f39ERkYCAD777DMkJCSgdOnS6N27N7Zt24Z3794BAJo1awYvLy+ULl0aXbt2xbp16/D69etsnT+vMNzoUJs24uvBg6KZlIjInNjbA3Fx8tx0OVFy/vz5tR6PHDkS27Ztw/Tp03H06FFERESgSpUqSEpKyvA47y8foFAooFKpsrW/pMvrbVng6emJa9euYcmSJbCzs0P//v3RoEEDvH37Fo6Ojjh37hx+++03uLu7Y8KECfD19TXIhVMZbnSoUiWgdGkgMRE4cEDuaoiI8pZCAeTPL89NnxMlHz9+HN27d0fbtm1RpUoVuLm54c6dO/o7YRqcnZ3h6uqK06dPa7YplUqcO3cuW8fx8fHB8ePHtbYdP34cFStW1Dy2s7NDYGAgFixYgMOHD+PEiRO4dOkSAMDKygoBAQGYNWsWLl68iDt37uCPP/7IxTvTDw4F1yGFQrTezJ8P7NwJtG0rd0VERJRbZcuWxdatWxEYGAiFQoHx48dn2AKjL4MGDcKMGTNQpkwZVKhQAQsXLsSLFy+ytWTB119/jQ4dOqB69eoICAjArl27sHXrVs3or5CQECiVStSpUwf29vZYu3Yt7Ozs4OXlhd27d+PWrVto0KABChYsiL1790KlUqF8+fL6ess5xpYbHVNfmtq9G8jlul9ERGQAfvjhBxQsWBD16tVDYGAgmjdvjho1auR5HaNGjULnzp3RrVs31K1bFw4ODmjevDlsbW2zfIygoCD8+OOPmDNnDipVqoTly5dj9erVaNSoEQCgQIECWLlyJfz9/VG1alUcPHgQu3btgouLCwoUKICtW7eiSZMm8PHxwbJly/Dbb7+hUqVKenrHOaeQ8vqCnsxiY2Ph7OyMmJgYODk56fz4b9+K+W5evACOHgU+/FDnpyAikt2bN29w+/ZtlCpVKlsfrqQ7KpUKPj4+6NChA6ZOnSp3OTqR0e9Vdj6/2XKjY/nyAa1aifsp5msiIiLKlbt372LlypW4fv06Ll26hH79+uH27dv4/PPP5S7N4DDc6AGHhBMRka5ZWFggJCQEtWrVgr+/Py5duoSDBw/Cx8dH7tIMDjsU60Hz5qIF5/p14No1wAD7WhERkZHx9PRMNdKJ0saWGz1wcgKaNBH32XpDRESUtxhu9EQ9aor9boiIiPIWw42eBAaKr3/9BWQw2zYRERHpGMONnnh6AjVqiCXd9uyRuxoiIiLzwXCjR7w0RURElPcYbvRIPST8wAEgIUHeWoiIiMwFw40e+foCJUoAr18D4eFyV0NERLrSqFEjDB06VPO4ZMmSmD9/foavUSgU2L59e67PravjZGTSpEmoVq2aXs+hTww3eqReSBPgkHAiIkMQGBiIFi1apPnc0aNHoVAocPHixWwf9/Tp0+jTp09uy9OSXsB4+PAhWrZsqdNzmRqGGz1Th5tduwAZFpElIqIUevbsibCwMPz333+pnlu9ejVq1qyJqlWrZvu4RYoUgb29vS5KzJSbmxtsbGzy5FzGiuFGzxo2FJP6PXoEnDoldzVERObt448/RpEiRRASEqK1PS4uDps3b0bPnj3x7NkzdO7cGcWKFYO9vT2qVKmC3377LcPjvn9Z6saNG2jQoAFsbW1RsWJFhIWFpXrNqFGjUK5cOdjb26N06dIYP3483r59CwAICQnB5MmTceHCBSgUCigUCk3N71+WunTpEpo0aQI7Ozu4uLigT58+iIuL0zzfvXt3BAUFYc6cOXB3d4eLiwsGDBigOVdWqFQqTJkyBcWLF4eNjQ2qVauG0NBQzfNJSUkYOHAg3N3dYWtrCy8vL8yYMQMAIEkSJk2ahBIlSsDGxgYeHh4YPHhwls+dE1x+Qc+srYGWLYGNG8WoqQ8+kLsiIiI9kSTRyVAO9vaiL0AmrKys0K1bN4SEhGDs2LFQ/P81mzdvhlKpROfOnREXFwc/Pz+MGjUKTk5O2LNnD7p27Qpvb2/Url0703OoVCq0a9cOrq6uOHnyJGJiYrT656g5OjoiJCQEHh4euHTpEnr37g1HR0d888036NixIy5fvozQ0FAcPHgQAODs7JzqGPHx8WjevDnq1q2L06dP4/Hjx+jVqxcGDhyoFeAOHToEd3d3HDp0CDdv3kTHjh1RrVo19O7dO9P3AwA//vgj5s6di+XLl6N69er4+eef0aZNG/zzzz8oW7YsFixYgJ07d2LTpk0oUaIEoqKiEBUVBQDYsmUL5s2bhw0bNqBSpUqIjo7GhQsXsnTeHJPMTExMjARAiomJybNzrlsnSYAkVaqUZ6ckItKrhIQE6cqVK1JCQkLyxrg48Z+dHLe4uCzXHhkZKQGQDh06pNlWv3596Ysvvkj3Na1bt5ZGjBihedywYUNpyJAhmsdeXl7SvHnzJEmSpP3790tWVlbS/fv3Nc/v27dPAiBt27Yt3XPMnj1b8vPz0zyeOHGi5Ovrm2q/lMdZsWKFVLBgQSkuxfvfs2ePZGFhIUVHR0uSJEnBwcGSl5eX9O7dO80+n332mdSxY8d0a3n/3B4eHtJ3332ntU+tWrWk/v37S5IkSYMGDZKaNGkiqVSqVMeaO3euVK5cOSkpKSnd86ml+Xv1f9n5/OZlqTzQsiVgZQX88w/w779yV0NEZN4qVKiAevXq4eeffwYA3Lx5E0ePHkXPnj0BAEqlElOnTkWVKlVQqFAhODg4YP/+/bh3716Wjh8ZGQlPT094eHhottWtWzfVfhs3boS/vz/c3Nzg4OCAcePGZfkcKc/l6+uL/Pnza7b5+/tDpVLh2rVrmm2VKlWCpaWl5rG7uzseZ3H6/NjYWDx48AD+/v5a2/39/REZGQlAXPqKiIhA+fLlMXjwYBw4cECz32effYaEhASULl0avXv3xrZt2/Du3btsvc/sYrjJAwULAg0aiPuc0I+ITJa9PRAXJ88tm515e/bsiS1btuDVq1dYvXo1vL290bBhQwDA7Nmz8eOPP2LUqFE4dOgQIiIi0Lx5cyQlJensW3XixAl06dIFrVq1wu7du3H+/HmMHTtWp+dIKV++fFqPFQoFVDoc5VKjRg3cvn0bU6dORUJCAjp06ID27dsDEKuZX7t2DUuWLIGdnR369++PBg0aZKvPT3Yx3OQR9YR+HBJORCZLoQDy55fnloX+Nil16NABFhYWWL9+PX755Rd8+eWXmv43x48fxyeffIIvvvgCvr6+KF26NK5fv57lY/v4+CAqKgoPHz7UbPv777+19vnrr7/g5eWFsWPHombNmihbtizu3r2rtY+1tTWUSmWm57pw4QLi4+M1244fPw4LCwuUL18+yzVnxMnJCR4eHjh+/LjW9uPHj6NixYpa+3Xs2BErV67Exo0bsWXLFjx//hwAYGdnh8DAQCxYsACHDx/GiRMncOnSJZ3UlxaGmzyiXkjz2DHg2TN5ayEiMncODg7o2LEjxowZg4cPH6J79+6a58qWLYuwsDD89ddfiIyMRN++ffHo0aMsHzsgIADlypVDcHAwLly4gKNHj2Ls2LFa+5QtWxb37t3Dhg0b8O+//2LBggXYtm2b1j4lS5bE7du3ERERgadPnyIxMTHVubp06QJbW1sEBwfj8uXLOHToEAYNGoSuXbvC1dU1e9+UDHz99deYOXMmNm7ciGvXrmH06NGIiIjAkCFDAAA//PADfvvtN1y9ehXXr1/H5s2b4ebmhgIFCiAkJASrVq3C5cuXcevWLaxduxZ2dnbw8vLSWX3vY7jJI6VKAVWqAEolsHev3NUQEVHPnj3x4sULNG/eXKt/zLhx41CjRg00b94cjRo1gpubG4KCgrJ8XAsLC2zbtg0JCQmoXbs2evXqhe+++05rnzZt2mDYsGEYOHAgqlWrhr/++gvjx4/X2ufTTz9FixYt0LhxYxQpUiTN4ej29vbYv38/nj9/jlq1aqF9+/Zo2rQpFi1alL1vRiYGDx6M4cOHY8SIEahSpQpCQ0Oxc+dOlC1bFoAY+TVr1izUrFkTtWrVwp07d7B3715YWFigQIECWLlyJfz9/VG1alUcPHgQu3btgouLi05rTEkhSZKkt6MboNjYWDg7OyMmJgZOTk55eu7x44Fp04D27YHNm/P01EREOvXmzRvcvn0bpUqVgq2trdzlkInI6PcqO5/fbLnJQ+rZikNDgTRaF4mIiEgHGG7ykJ8f4OEhOvYfOiR3NURERKaJ4SYPWVgkdyzmkHAiIiL9YLjJY+oh4Tt3imk1iYiISLcYbvJY48ZiSob794Fz5+Suhogod8xsTArpma5+n2QNN3/++ScCAwPh4eGRapXT9Bw+fBg1atSAjY0NypQpk2plV0Nnaws0by7uc0I/IjJW6hlvX8u1UCaZJPUMzSmXisgJWVcFj4+Ph6+vL7788ku0a9cu0/1v376N1q1b46uvvsK6desQHh6OXr16wd3dHc3VicEIfPIJsHWruDQ1ZYrc1RARZZ+lpSUKFCigWZ/I3t5eM8MvUU6oVCo8efIE9vb2sLLKXTwxmHluFAoFtm3bluFESaNGjcKePXtw+fJlzbZOnTrh5cuXCA0NzdJ55JznRu3pU8DVFVCpgDt3AD1O0khEpDeSJCE6OhovX76UuxQyERYWFihVqhSsra1TPZedz29ZW26y68SJEwgICNDa1rx5cwwdOjTd1yQmJmpNWR0bG6uv8rKscGHgww+BP/8UrTeDBsldERFR9ikUCri7u6No0aJ6XQSRzIe1tTUsLHLfY8aowk10dHSqtTJcXV0RGxuLhIQE2NnZpXrNjBkzMHny5LwqMcvatGG4ISLTYGlpmes+EkS6ZPKjpcaMGYOYmBjNLSoqSu6SACTPVnz4MMAWXSIiIt0xqnDj5uaWamXWR48ewcnJKc1WGwCwsbGBk5OT1s0QlC0L+PgA796J5RiIiIhIN4wq3NStWxfh4eFa28LCwlC3bl2ZKsoddesNh4QTERHpjqzhJi4uDhEREYiIiAAghnpHRETg3r17AMQlpW7dumn2/+qrr3Dr1i188803uHr1KpYsWYJNmzZh2LBhcpSfa+rZivftA/4/tJ+IiIhySdZwc+bMGVSvXh3Vq1cHAAwfPhzVq1fHhAkTAAAPHz7UBB0AKFWqFPbs2YOwsDD4+vpi7ty5+Omnn4xqjpuUatcGihYFYmKAo0flroaIiMg0GMw8N3nFEOa5SalXL2DVKjFiasECuashIiIyTNn5/DaqPjemSN3vhgtpEhER6QbDjcwCAgA7O+DuXeDiRbmrISIiMn4MNzKztweaNRP3d+6UtxYiIiJTwHBjADgknIiISHcYbgzAxx8DCgVw9izw339yV0NERGTcGG4MgKsr8MEH4v7u3fLWQkREZOwYbgyEekI/XpoiIiLKHYYbA6Hud/PHH8CrV/LWQkREZMwYbgxEhQpiMc2kJGD/frmrISIiMl4MNwZCodCe0I+IiIhyhuHGgKjDzZ49wLt38tZCRERkrBhuDEi9eoCLC/D8OXD8uNzVEBERGSeGGwNiZQW0bi3u89IUERFRzjDcGJiUQ8K5kCYREVH2MdwYmI8+AmxsgH//BSIj5a6GiIjI+DDcGBgHB6BpU3GfE/oRERFlH8ONAeKQcCIiopxjuDFAgYHi68mTQHS0vLUQEREZG4YbA+ThAdSqJToUcyFNIiKi7GG4MVC8NEVERJQzDDcGSj0kPCwMiI+XtxYiIiJjwnBjoCpXBkqWBN68AQ4elLsaIiIi48FwY6AUCu0J/YiIiChrGG4MmLrfze7dgFIpby1ERETGguHGgNWvDxQoADx5Avz9t9zVEBERGQeGGwOWLx/QqpW4z1FTREREWcNwY+A4JJyIiCh7GG4MXIsWogXn6lXg+nW5qyEiIjJ8DDcGztkZaNRI3GfrDRERUeYYbowAh4QTERFlHcONEVAvpPnXX2LkFBEREaWP4cYIlCgBVKsGqFTA3r1yV0NERGTYGG6MBC9NERERZQ3DjZFQDwnfv1+sN0VERERpY7gxEtWrA8WLA69fA+HhcldDRERkuBhujIRCwQn9iIiIsoLhxoio+93s3Ck6FxMREVFqDDdGpGFDwNERiI4GzpyRuxoiIiLDxHBjRGxsxHIMAC9NERERpYfhxshwSDgREVHGGG6MTMuWgKUlcPkycOuW3NUQEREZHoYbI1OoENCggbjPS1NERESpMdwYIQ4JJyIiSh/DjRFSh5s//wSeP5e3FiIiIkPDcGOESpcGKlcGlEpg3z65qyEiIjIsDDdGipemiIiI0sZwY6TUQ8L37QMSE+WthYiIyJAw3BipmjUBd3fg1SvgyBG5qyEiIjIcDDdGysICCAwU9zmhHxERUTKGGyOWst+NJMlbCxERkaFguDFiTZsC9vbAf/8B58/LXQ0REZFhYLgxYra2QPPm4j5HTREREQkMN0aOQ8KJiIi0MdwYudatRefi8+eBe/fkroaIiEh+soebxYsXo2TJkrC1tUWdOnVw6tSpDPefP38+ypcvDzs7O3h6emLYsGF48+ZNHlWbBXncs7dIEaBePXF/1648PTUREZFBkjXcbNy4EcOHD8fEiRNx7tw5+Pr6onnz5nj8+HGa+69fvx6jR4/GxIkTERkZiVWrVmHjxo349ttv87jyNDx+DHzxBTB1ap6fWj2hH4eEExERAQpJkm8QcZ06dVCrVi0sWrQIAKBSqeDp6YlBgwZh9OjRqfYfOHAgIiMjER4ertk2YsQInDx5EseOHcvSOWNjY+Hs7IyYmBg4OTnp5o0AwMaNQKdOgLU1cOECUKGC7o6dievXgfLlgXz5gCdPAGfnPDs1ERFRnsjO57dsLTdJSUk4e/YsAgICkouxsEBAQABOnDiR5mvq1auHs2fPai5d3bp1C3v37kWrVq3SPU9iYiJiY2O1bnrRoQPQsiWQlAT07QuoVPo5TxrKlRNZ6u1bIDQ0z05LRERkkGQLN0+fPoVSqYSrq6vWdldXV0RHR6f5ms8//xxTpkzBhx9+iHz58sHb2xuNGjXK8LLUjBkz4OzsrLl5enrq9H1oKBTAkiVi4pk//wRWr9bPedLBUVNERESC7B2Ks+Pw4cOYPn06lixZgnPnzmHr1q3Ys2cPpmbQz2XMmDGIiYnR3KKiovRXYMmSwJQp4v7IkcCjR/o713vU4WbvXtGCQ0REZK5kCzeFCxeGpaUlHr0XAB49egQ3N7c0XzN+/Hh07doVvXr1QpUqVdC2bVtMnz4dM2bMgCqdy0A2NjZwcnLSuunVkCFA9erAy5fAsGH6PVcKH3wgRk69fAkcPZpnpyUiIjI4soUba2tr+Pn5aXUOVqlUCA8PR926ddN8zevXr2FhoV2ypaUlAEDGftHarKyAlSvF5DO//ZZnnWAsLYGPPxb3eWmKiIjMmayXpYYPH46VK1dizZo1iIyMRL9+/RAfH48ePXoAALp164YxY8Zo9g8MDMTSpUuxYcMG3L59G2FhYRg/fjwCAwM1Iccg+PmJFhwA6NcPiI/Pk9OmHBJuKFmPiIgor1nJefKOHTviyZMnmDBhAqKjo1GtWjWEhoZqOhnfu3dPq6Vm3LhxUCgUGDduHO7fv48iRYogMDAQ3333nVxvIX1TpgBbtgB37gCTJgGzZ+v9lAEBYr2pO3eAy5eBKlX0fkoiIiKDI+s8N3LQ2zw3admzR1wrsrQETp8WfXH0rE0bMVPx1KnAuHF6Px0REVGeMIp5bsxC69Zi/hulEujdW3zVMw4JJyIic8dwo28//iimDD57Fli4UO+n+/hjMeXO6dPAgwd6Px0REZHBYbjRNzc3YNYscX/cOL0v3e3mBtSpI+5zIU0iIjJHDDd5oVcv4MMPxaip/v31PpSJl6aIiMicMdzkBQsLYMUKsbLlnj3A77/r9XTqIeHh4UBcnF5PRUREZHAYbvKKjw+gnrNn8GAxlbAeT+XtDSQmAgcO6O00REREBonhJi+NGQOULw9ERwOjR+vtNAqF9oR+RERE5oThJi/Z2gLLl4v7y5cDx47p7VTqfjd79gDv3untNERERAaH4SavNWwI9Owp7vfpI64d6YG/P1CoEPDsGXDihF5OQUREZJAYbuQwaxZQtCgQGZk8TFzHrKzEHIIAL00REZF5YbiRQ6FCwPz54v60acC1a3o5jfrSFBfSJCIic8JwI5dOnYAWLYCkJOCrr/SSPpo3B6ytgZs3gatXdX54IiIig8RwIxeFAliyBLCzAw4fBkJCdH4KR0egSRNxnxP6ERGRuWC4kVOpUsCUKeL+iBHA48c6PwWHhBMRkblhuJHb0KFAtWrAixfAsGE6P/zHH4uvf/8NPHqk88MTEREZHIYbuVlZiaUZLCyA9euB/ft1evjixQE/P9GlZ88enR6aiIjIIDHcGIJatcSSDADQrx/w+rVOD89LU0REZE4YbgzF1KmApydw+zYwebJOD60eEh4WpvPcREREZHAYbgyFgwOweLG4P3cuEBGhs0NXrQp4eQEJCcDBgzo7LBERkUFiuDEkgYFA+/aAUgn07i2+6oBCkdx6wyHhRERk6hhuDM2CBYCzM3DmTHJLjg6o+93s2qWzzERERGSQGG4Mjbs7MHOmuD92LBAVpZPDNmggMtPjx8CpUzo5JBERkUFiuDFEvXuLZb3j4oABA3SyNEO+fEDLluI+L00REZEpY7gxRBYWwPLlIpHs2gVs3aqTw3JIOBERmQOGG0NVqRIwerS4P2gQEBOT60O2aCHmDIyMBG7cyPXhiIiIDBLDjSH79lugXDng4UNgzJhcH65AAaBRI3Gfl6aIiMhUMdwYMltbcXkKAJYuBY4fz/UhOSSciIhMHcONoWvUCOjRQ9zv0wdISsrV4dTh5tgx4OnT3JVGRERkiBhujMHs2UCRIsCVK+J+Lnh5Ab6+gEoF7N2ro/qIiIgMCMONMXBxAebPF/enTgWuX8/V4XhpioiITBnDjbHo3Bn46CMgMRH46qtczX2jHhIeGgq8eaOj+oiIiAwEw42xUChEp2I7O+DQIWDNmhwfqkYNoFgxID5eHIqIiMiUMNwYk9KlgUmTxP0RI4AnT3J0mJQLaXJCPyIiMjUMN8Zm2DDRI/j5c2D48BwfRh1udu0SnYuJiIhMBcONscmXD1i5UjS/rF0LhIXl6DCNGwMODsCDB8C5czqukYiISEYMN8aoVi2xJAMgOhe/fp3tQ9jYiOUYAF6aIiIi08JwY6ymTQOKFwdu3QKmTMnRITgknIiITBHDjbFydAQWLxb358wBLlzI9iFatwYsLYGLF4Hbt3VcHxERkUwYboxZmzbAp58CSqVYmkGpzNbLCxUCPvxQ3N+1Sw/1ERERyYDhxtgtWAA4OQGnTgFLlmT75eoJ/djvhoiITAXDjbHz8AC+/17c//ZbICoqWy9X97s5cgR48ULHtREREcmA4cYU9O0L1KsHxMUBAwdma2kGb2+gUiVxRSs0VI81EhER5RGGG1NgYQGsWCHmwNm5E9i2LVsv52zFRERkShhuTEWlSsA334j7gwYBMTFZfqk63OzbByQl6aE2IiKiPMRwY0rGjQPKlhXTDn/7bZZfVrs24OoKxMaKvjdERETGjOHGlNjaAsuWiftLlwInTmTpZRYWQGCguM8J/YiIyNjlKNxERUXhv//+0zw+deoUhg4dihUrVuisMMqhJk2A7t1Fp+I+fbJ8nSnlkPBs9EcmIiIyODkKN59//jkOHToEAIiOjkazZs1w6tQpjB07FlNyuBQA6dCcOUDhwsDly+J+FjRtCtjbi5HkOZjsmIiIyGDkKNxcvnwZtWvXBgBs2rQJlStXxl9//YV169YhJCREl/VRTri4APPmiftTpgA3bmT6Ejs74KOPxH1emiIiImOWo3Dz9u1b2NjYAAAOHjyINv8fblOhQgU8fPhQd9VRznXpAjRrBiQmipXDs3CtiUPCiYjIFOQo3FSqVAnLli3D0aNHERYWhhYtWgAAHjx4ABcXF50WSDmkUIhOxba2wB9/AL/+mulLWrcWLzt3DkjRpYqIiMio5CjczJw5E8uXL0ejRo3QuXNn+Pr6AgB27typuVxFBsDbG5g0SdwfPhx48iTD3YsWFRMdA7w0RURExkshSTkbG6NUKhEbG4uCBQtqtt25cwf29vYoWrSozgrUtdjYWDg7OyMmJgZOTk5yl6N/b98CNWsCFy8CXbsCv/yS4e6zZgGjRon+N/v351GNREREmcjO53eOWm4SEhKQmJioCTZ3797F/Pnzce3atWwHm8WLF6NkyZKwtbVFnTp1cOrUqQz3f/nyJQYMGAB3d3fY2NigXLly2Lt3b07ehnnIlw9YuVJcb/r1V+DgwQx3DwoSXw8cyPYqDkRERAYhR+Hmk08+wS//bwF4+fIl6tSpg7lz5yIoKAhLly7N8nE2btyI4cOHY+LEiTh37hx8fX3RvHlzPH78OM39k5KS0KxZM9y5cwe///47rl27hpUrV6JYsWI5eRvmo3ZtsaAmIBbZfP063V3LlQMGDxb3v/gCiIjQf3lEREQ6JeWAi4uLdPnyZUmSJGnlypVS1apVJaVSKW3atEmqUKFClo9Tu3ZtacCAAZrHSqVS8vDwkGbMmJHm/kuXLpVKly4tJSUl5aRsSZIkKSYmRgIgxcTE5PgYRikmRpKKFZMkQJJGj85w17dvJalZM7Grp6ckRUfnUY1ERETpyM7nd45abl6/fg1HR0cAwIEDB9CuXTtYWFjggw8+wN27d7N0jKSkJJw9exYBAQGabRYWFggICMCJdJYN2LlzJ+rWrYsBAwbA1dUVlStXxvTp06FUKtM9T2JiImJjY7VuZsnJCVi8WNyfM0f0wUmHlRWwcaNYpioqCmjXTowoJyIiMgY5CjdlypTB9u3bERUVhf379+Oj/8/+9vjx4yx30n369CmUSiVcXV21tru6uiI6OjrN19y6dQu///47lEol9u7di/Hjx2Pu3LmYNm1auueZMWMGnJ2dNTdPT88svksT9MknIqm8eyeWZsggFBYsCOzaBTg7A3/9leWpcoiIiGSXo3AzYcIEjBw5EiVLlkTt2rVRt25dAKIVp3r16jotMCWVSoWiRYtixYoV8PPzQ8eOHTF27FgsUy8WmYYxY8YgJiZGc4uKitJbfUZhwQLA0RE4eTJ5kc10lC8PbNokFtYMCUme9JiIiMiQ5SjctG/fHvfu3cOZM2ewP8V44aZNm2JeFj8BCxcuDEtLSzx69Ehr+6NHj+Dm5pbma9zd3VGuXDlYWlpqtvn4+CA6OhpJ6SwQaWNjAycnJ62bWStWDPj+e3F/zJhMZ+v76CPghx/E/a+/BjgwjYiIDF2Owg0AuLm5oXr16njw4IFmhfDatWujQoUKWXq9tbU1/Pz8EB4ertmmUqkQHh6uaQl6n7+/P27evAmVSqXZdv36dbi7u8Pa2jqnb8X8fPUVULcu8OoVMGhQprsPHgz06gWoVEDnzsCVK3lQIxERUQ7lKNyoVCpMmTIFzs7O8PLygpeXFwoUKICpU6dqBY/MDB8+HCtXrsSaNWsQGRmJfv36IT4+Hj169AAAdOvWDWPGjNHs369fPzx//hxDhgzB9evXsWfPHkyfPh0DBgzIydswXxYWwIoVoufw9u2ZTmijUIi+yPXrA7GxYg2qZ8/yplQiIqLsssrJi8aOHYtVq1bh+++/h7+/PwDg2LFjmDRpEt68eYPvvvsuS8fp2LEjnjx5ggkTJiA6OhrVqlVDaGioppPxvXv3YGGRnL88PT2xf/9+DBs2DFWrVkWxYsUwZMgQjBo1Kidvw7xVrgx88w0wfbpovWnaVIyoSoe1NbBli5gy599/gQ4dgNBQMUcgERGRIcnR8gseHh5YtmyZZjVwtR07dqB///64f/++zgrUNbNbfiEjCQlA1arAzZtikr+FCzN9yaVLYv2puDigf//k0eVERET6pPflF54/f55m35oKFSrg+fPnOTkkycHOLnnE1OLFwN9/Z/qSKlWAdevEpaolS8SNiIjIkOQo3Pj6+mLRokWpti9atAhVq1bNdVGUh5o2BYKDxSQ2vXuLhTYz0aaNuJoFiM7Gf/yh5xqJiIiyIUeXpY4cOYLWrVujRIkSmpFNJ06cQFRUFPbu3Yv69evrvFBd4WWpNDx9ClSoIHoJz5gBjB6d6UskSSwyvm6dmPDv1CmgTJk8qJWIiMyS3i9LNWzYENevX0fbtm3x8uVLvHz5Eu3atcM///yDX3/9NUdFk4wKF06eoW/yZNFjOBMKBfDTT6KD8YsXojUnJkbPdRIREWVBjlpu0nPhwgXUqFEjw7We5MaWm3RIEtCsGRAeDgQEAAcOiASTiYcPgVq1gPv3gZYtxZINKeZYJCIi0gm9t9yQCVIoROdiW1vg4EEgiy1w7u7Ajh2ib/K+fQBH5RMRkdwYbihZmTLAxInifvfuQPv2Yg2qTPj5ibWnAGDu3OT7REREcmC4IW0jRogZ+iRJzNr3wQdAw4bA7t1i/YV0dOgATJgg7vftCxw/nkf1EhERvSdbfW7atWuX4fMvX77EkSNH2OfGFFy+DMyZA6xfnzw8vGJFYORI4PPPARubVC9RqUTI2bIFKFpUjKDy8srjuomIyCRl5/M7W+FGveZTZlavXp3VQ+Y5hpts+u8/4McfgeXLxUKbAODhAQwZIpponJ21do+PB/z9gQsXAF9f4NgxwMFBhrqJiMik6C3cmAKGmxyKiREB58cfgQcPxDZHRxFwhgwBihfX7HrvnhhB9fgx0LYt8PvvYq1OIiKinOJoKdI9Z2ex0OatW8DPP4tLVK9eiUtXpUqJDsiXLwMASpQQC41bW4uvkybJWjkREZkZhhvKHhsboEcPsYLm7t2is/G7d8CaNWLhqVatgEOHUK+uhBUrxEumTgU2bpS3bCIiMh8MN5QzFhZA69bA4cNiuHj79mKunH37gCZNgNq1EWy3Cd8MfwdANOycOSNrxUREZCYYbij3atcGNm8Grl8H+vUTEwGeOQN07Ijvt5fH4oqLoXjzGp98ktxdh4iISF8Ybkh3ypQBliwRPYonTABcXKC4dQv9rwzEfcsS6P1gErq3foKEBLkLJSIyQC9fisv8lGscLUX6Ex8PrF4N/PADcPs2AOA17HC8bA8E7B0ORRlvmQskIpLJmzfAuXPisv7ff4vbvXuAlZWYIMzbO/WtdGkgf365K5cNh4JngOFGBu/eAVu3Inb8LDhdPwsAUCksYPFpOzECq1YtmQskItIjSRIjTf/+OznMREQkT5CaHW5uaQcfb2+gcOEsLXhsrBhuMsBwIyNJws7hh2E1fzZaYV/y9oYNRchp2dKk/2ESkZmIiQFOn05ukTl5Enj6NPV+RYsCdesCdeqIpW5q1ABiY4F//0379vJlxud1chKtO2kFH09PwNJSL283rzDcZIDhRn4DBwJHFl/CaKs5+BzroVBfY65UKXl5B2treYskIsoKpRK4ciU5yPz9NxAZKVprUrK2FuHlgw+Sw4yXV/b+oHv+PP3gc/9+xq/Nlw8oWTL9y112dtl+63mN4SYDDDfye/tWNNKEhwMfFItCeJsfYb92RfLyDsWKiVmP+/RJtbwDEZGsHj3S7idz+jQQF5d6v1KlRIBRh5lq1dJck09nEhJE38a0gs/t25lfAvPwSP9yV6FCBtGqznCTAYYbw/D8ufj3fvMm8OGHQPiWl7Be/f/lHR4+FDs5OSUv71CsmLwFE5H5SUwEzp/XDjN37qTez8FBTImRMswULZrn5aZLqRTrBKbX6hMbm/HrnZ3TDz7Fi+fZ+joMNxlguDEckZHi/4HYWODLL4GffgIUSYliJfLZs8UOgGhO7dJFXLKqVEneoonINEmSCC4pO/2ePw8kJWnvp1CI5WfUQeaDDwAfH+PtzyJJwLNn6Qcf9R+b6bG2Fq1U74eecuXETYcYbjLAcGNY9u0DPv4YUKmAefOAoUP//4RKBezdC8yaBRw9mvyCVq1E5+MGDQyimZSIjNSrV8mdftVh5vHj1PsVKZLcR+aDD8ToTnP67Hj9Woz0Siv43LmT/rw8lSuLZXp0iOEmAww3hmfePGD4cNGyuXcv0Lz5ezucPClacrZuTe6kV6uWCDlt2xrvX0xElDdUKtESnHL00uXLqTv95ssHVK+uHWZKleIfUul59w6Iiko7+FSuDKxdq9PTMdxkgOHG8EgS0LOnmO/P2Vn831OhQho73rghJgRcvVpcCwdE8+eIEWLxKiPo7U9EeeDJk9SdftPqV+LlpT16qXp1sXwMGSSGmwww3BimxESgaVPg+HGxisPJk6KDfpoePwYWLQIWLxY9kwExedXAgcCAAeI+EZkHlUrM9HviRHKYuXUr9X7584sWX3WYqVMHcHfP+3opxxhuMsBwY7gePxb/99y7BwQEiP44VlYZvCA+Hvj5Z9Gaox7BYGcneicPHy7mbiAi03T7NrBmjbilNYLJx0d79FKlSpn8h0KGjuEmAww3hu3CBcDfX+SWQYOABQuy8KJ374AtW0Tn43PnxDYLC6B9e2DYMPEfG6+ZExm/uDjxbz0kBDh8OHm7oyNQv752p98CBWQqkvSF4SYDDDeGb/t20U8YAJYtE1PdZIkkAYcOiZCzf3/ydh8fIDgY+OILzpdDZGwkSYyYXL0a2LxZ/OUDiD9YAgJEf7ugIMDeXs4qKQ8w3GSA4cY4fPcdMG6caEUOCwMaNcrmAS5eBObOBTZtEqvvAqI1JyBABB3+Z0hk2O7cAX75RVx2StmHpkwZEWi6dRPrJZHZYLjJAMONcZAkscTUhg2iY/Hp0znsQhMTI/7aW7MGOHYsebujI9Chgwg6H37Iy1b6lpAgUuqTJ0CbNmLuEKL3xceLKR9CQoA//kje7ugIdOwoQk29evz3aqYYbjLAcGM8EhLEXH1nzoi+gH/9lcu5s/79V/wl+Msv2h0QS5cWfwV26ybmtCDdePEC2L1bXGcMDRWTgQFiRtN27cT1xoYN+UFl7iRJDJNcvVq0tKrXaVIogCZNRKBp144trcRwkxGGG+Py4IHoG/jggZjJePt2HczZp1KJa/hr1ohWnZSL3jVoIFpzPvtM/LVI2XP/vvghbdsGHDmiPXuppyfg4gJERCRvK1dOLJAaHMwh/Obm3j3xh0ZIiPjDQ83bWwSarl3FPDRE/8dwkwGGG+Nz+rTIHG/eiEmJZ87U4cHj48UH8Zo1Yply9T8He3vx12JwMNC4MWdBzsjVq+J7uG2b+GGlVKmS6B0eFATUqCH+Gj97FlixQqwhpg6W1tZidFvfvmLUC1tzTNPr1+L3JCRE+9+bg4O4TNy9Oy8TU7oYbjLAcGOcfvtN9MEBRA7p1k0PJ4mKEtOFh4QA168nby9eXPwVGRwMlC+vhxMbGZVKXCtUB5pr15KfUyjEUFx1oClbNv3jvHolAs7y5WKBQrUKFZJbc9KdyZGMhiSJa8ohIcDGjeLnrta4sQg0n34qJtkjygDDTQYYbozXuHFiFJW1tZjiom5dPZ1IksQUyWvWiB7NL18mP1enjvjQ7dQJKFhQTwUYoLdvxTd9+3Zgxw5x+UktXz4xvXRQkOgsnJNZX8+cESHnt9+Sh/ra2IjLg337ismP+Ne8cYmKAn79VYSaGzeSt5cqlTzaqWRJmYojY8RwkwGGG+OlUok/8LZvB1xdxRUQvY8EffMG2LVLBJ3QUECpFNutrYFPPhFBp3lz05z5ND5evOft20XH4JQhz8FBrNAeFCS+Ojvr5pyxscC6dSLoXLiQvL1iRdGa062beYVKY5OQIH5fVq8GDh5MvuyUP78Iqt27i8uOFhZyVklGiuEmAww3xi0uTvwRf/GiWOPu6NE8bM2OjhaXUUJCgEuXkre7ugJduoigU7VqHhWjJ0+fijC3fTtw4EDyHEGAGL79ySci0DRtqt8FBiVJpNfly0XrmXqkla2t6JvRt69oumNrjvwkSaznFBIiflYpF6hs2BDo0UP8VeLgIFuJZBoYbjLAcGP87t4VI6iePBH/Z27alMd/CEqSGPGzZo0IO0+eJD9XrZoIOZ9/DhQtmodF5cLdu8kjnI4eFU1kaqVKJfefqVdPno7VMTGiL9Ty5dqhsnJl0ZrTtSun2pfD/fvJl51S9rsqWVL8G+jWjeu7kU4x3GSA4cY0HDsmpsB4+xaYOBGYNEmmQt6+FSt8rlkjWjzevhXbrayAli3Ff/Iffyz6jxgKSQIuX04ONCk78wKAr68ING3bAlWqGE7riLov1PLlomNqQoLYbmcnJnjr25friOnbmzeiz9Xq1WJSRnUQtrcXo9169BBDG3nZifSA4SYDDDem4+efgZ49xf1Nm8QlfVk9eyaa5des0R4SXagQ0LmzCDo1a8rz4atSASdOJAealPOKWFiI4bdBQeJmDBMZvnyZ3Jpz+XLy9qpVRWvOF1/orh+QuZMk4NSp5MtOKfteNWgg+tG0b895oUjvGG4ywHBjWkaMAH74QfzxfuyYmErFIFy5IiYo+/VXMQOhWl4u4pmYKKawV49wevQo+TkbG6BZMxFmAgON5xLa+yRJhLbly7XXEbO3FyPa+vYV1zDZmpN9Dx4kT40QGZm8vUQJ8TscHCwm3CPKIww3GWC4MS1KpbjqExoqpqM5dSpnI5H1RqkUo0bWrBEtJvpexDM2Vlwm274d2LNHe04RJyfxzQoKAlq0ML2/tF+8EGFy+XIRLtWqVROtOV265HL9DjPw5g2wc6cINPv3J192srMTrTPdu4tVbHnZiWTAcJMBhhvTExMj5o27elV0uTh8WL8DeXIsvUU8nZySh8nmZD6XR4/EB9K2bWLW16Sk5Ofc3JIvNzVuLIawmzr1WkXLl4vvd2Ki2J4/v7g82LevuDxIgiSJeYZCQsQ8Qy9eJD/34Yfi9/KzzxgMSXYMNxlguDFNN26IYPPihRg8s2aNgV+JSG8RT2/v5EU8M5rg7Nat5BmC//oreT4RQMwKrO4QXLu2ef+V/fy5+B4vXy7Sr1qNGqI15/PPTa8FKzNJScB//4lRcqdPi38sKVu6ihdPvuyU0QzTRHmM4SYDDDemKzxczKenVIr1p775Ru6KskC9iGdICPD779qLeDZsKD5g2rcXc4RERCR3CE45JBoA/PySA42Pj4EnOxlIkvg+r1ghvs/q1hwHBxFw+vY1oA5buRQTI4LLvXvaX9X3Hz7UDsOAaOr89FPRSsO11MhAMdxkgOHGtC1eDAwcKD7bd+wQfWWNRnw8sHWr+Ev6jz+0F/EsXFh8MKlZWorwo77kpPepmk3Is2fie7x8ufYaYjVritaczp0Nd8I5lUpMJplReImJyfw4traiY3Dp0mKB2A4dOLqMDB7DTQYYbkybJAH9+wPLlonPpxMnxFxvRufePTFSZc2a5A9gOzvRNBUUJDoGu7jIWqLRkyTgyBHRmrNlS3JfJUdH0fm4b1/RGTkvvXkj1mRKL7xERSXPpZQRFxcRXry8kr+mvF+kCFv3yOgw3GSA4cb0vX0LfPSR6FhcvLjoI/nhh3JXlUPqZQiePRNzinDlZP148kQEyRUrtBd5rF1btOZ06pT7770kiTliUrayvH8/5XD99FhaimkE0gotJUqIm6G2PBHlAsNNBhhuzMOzZ2K1gOvXRX/aUaPELMbmMFiIckGSRCpevlxcIlS3kjg5ibmJ+vZNf/0wpVLMDZPe5aK7d7X7VKXH3j7t0KLe5uFhmgu1EmWC4SYDDDfmIyYGGDJE/EEOiIU2164VC0wTZerxY9HRe8UK7RmdP/hAdPJ+vxXmv/+Ad+8yP27RohlfMipUiJeMiNLAcJMBhhvzs2WLuLLw/LnoRzlzpuh0bM4jpCkbVCrg0CHRmrNtW8YBxspKdO5Oq9VFfcnIzi7vaicyIQw3GWC4MU8PHwJffilmMgbEygOrV+t/BQQyMY8eiV+cU6fEBInvt764uXEYNZGeZOfz2yD+dl28eDFKliwJW1tb1KlTB6dOncrS6zZs2ACFQoGgoCD9FkhGz90d2LsXWLRI/OEcFiYWvN64Ue7KyKi4ugKjR4v+OEuWiM5cnTuLDl7FijHYEBkI2cPNxo0bMXz4cEycOBHnzp2Dr68vmjdvjsePH2f4ujt37mDkyJGoX79+HlVKxk6hAAYMAM6dE1OavHghBsF88YX2QsdERGTcZA83P/zwA3r37o0ePXqgYsWKWLZsGezt7fHzzz+n+xqlUokuXbpg8uTJKF26dB5WS6agQgWxYsH48aLfzbp1YgDMoUNyV0ZERLoga7hJSkrC2bNnERAQoNlmYWGBgIAAnDhxIt3XTZkyBUWLFkXPnj0zPUdiYiJiY2O1bkT58gFTpoj1K729xdxoTZsCI0cmL9xNRETGSdZw8/TpUyiVSri6umptd3V1RXR0dJqvOXbsGFatWoWVK1dm6RwzZsyAs7Oz5ubJaeophbp1xZJNvXuLKU7mzhXztl28KHdlRESUU7JflsqOV69eoWvXrli5ciUKFy6cpdeMGTMGMTExmltUVJSeqyRj4+AgpjLZsUPMSn/pElCrFjB7tpiXjYiIjIus01wWLlwYlpaWePTelOOPHj2Cm5tbqv3//fdf3LlzB4EpVkNUqVQAACsrK1y7dg3e3t5ar7GxsYGNjY0eqidT06YNcPky0KsXsGuXWFV8zx4xCaCXl9zVERFRVsnacmNtbQ0/Pz+Eh4drtqlUKoSHh6Nu3bqp9q9QoQIuXbqEiIgIza1NmzZo3LgxIiIieMmJcq1oUdGCs3KlWEroyBHR2fjXX5MX6SYiIsMm+wIlw4cPR3BwMGrWrInatWtj/vz5iI+PR48ePQAA3bp1Q7FixTBjxgzY2tqi8ntLPBcoUAAAUm0nyimFQrTeNG4MdO0qVhbv1k205ixdysW4iYgMnex9bjp27Ig5c+ZgwoQJqFatGiIiIhAaGqrpZHzv3j08fPhQ5irJHHl7A3/+CUybJmbV37xZTPy3f7/clRERUUa4/AJRFpw5Iyb7u3ZNPB44UKxRZW8vb11ERObC6JZfIDJ0NWuKmY0HDhSPFy0C/PyAs2flrYuIiFJjuCHKInt7YOFCsfimuztw9SrwwQfislVGC0UTEVHeYrghyqbmzcVcOO3bi1AzfjzQoAHw779yV0ZERADDDVGOuLgAmzYBv/wCODmJEVW+vsBPP3HIOBGR3BhuiHJIoRBDxS9eFC038fFiGYegICCTRe2JiEiPGG6IcsnLC/jjD2DWLLEg586dYsj4rl1yV0ZEZJ4Yboh0wNIS+Ppr4PRpoHJl0XLTpg3Qpw8QFyd3dURE5oXhhkiHfH1FwBk+XDxeuRKoXh34+2956yIiMicMN0Q6ZmsLzJ0LhIcDnp7AzZuAvz8wYQLw9q3c1RERmT6GGyI9adJEdDbu0gVQqYCpU4F69ZJnOSYiIv1guCHSowIFgLVrgQ0bxP0zZ8RlqiVLOGSciEhfGG6I8kDHjmLiv6ZNgYQEYMAAoFUrgGvCEhHpHsMNUR4pXhw4cACYPx+wsRHLOFSpAmzdKndlRESmheGGKA9ZWABDhohFOKtVA549Az79FOjeHYiNlbs6IiLTwHBDJIOKFYGTJ4ExY0TgWbMGqFoVOHpU7sqIiIwfww2RTKytgenTgSNHgFKlgLt3gYYNgdGjgcREuasjIjJeDDdEMvvwQyAiAvjySzGCauZMoE4d4J9/5K6MiMg4MdwQGQAnJ2DVKtG5uHBh4MIFwM9PdD5WqeSujojIuDDcEBmQtm3FkPFWrcSlqWHDgGbNgKgouSsjIjIeDDdEBsbNDdi9G1i6FLC3FyuOV64M/Pgj8O6d3NURERk+hhsiA6RQAF99BZw/L/rfxMYCQ4eK2Y0PH5a7OiIiw8ZwQ2TAypUDjh8Hli8HXFyAy5eBxo2BTp2A//6TuzoiIsPEcENk4CwtgT59gOvXgf79xbw4GzcC5csDM2Zw2DgR0fsYboiMRKFCwOLFwNmzgL8/8Po18O23oj/O3r1yV0dEZDgYboiMTLVqYibjX38VnY9v3gRatwYCA4F//5W7OiIi+THcEBkhhQL44gvg2jVg5EjAykqMsKpYERg3DoiPl7tCIiL5MNwQGTEnJ2D2bDE3TrNmQFIS8N13gI8PsHmzmPGYiMjcMNwQmYAKFYD9+8UMx15eYtK/Dh2AgAAu40BE5ofhhshEKBRihuMrV4CJEwEbGzEBoK8vMHw4EBMjd4VERHmD4YbIxNjbA5MmAZGRQFAQoFQC8+aJoeNr1nCtKiIyfQw3RCaqVClg2zYgNFRMBvjoEdC9uxhGfvas3NUREekPww2RiWveXHQ4njkTyJ8f+PtvoFYtoG9f4OlTuasjItI9hhsiM2BtDXzzjRg6/vnnYhTVihWiRWfJEnHpiojIVDDcEJmRYsWAdeuAI0eAqlWBFy+AAQOAmjWBY8fkro6ISDcYbojMUIMGot/NokVAgQJARARQvz7QtSvw8KHc1RER5Q7DDZGZsrISrTbXrwO9e4uh5GvXiktVc+aICQGJiIwRww2RmStSRPS/OXkSqFMHiIsDvv5aXLY6cEDu6oiIso/hhogAiBFUf/0F/PyzCDzXromRVu3aAXfuyF0dEVHWMdwQkYaFBdCjh7hUNWQIYGkp5srx8QEmTwYSEuSukIgocww3RJRKgQLA/Pmio3GjRsCbN2LW44oVge3buSAnERk2hhsiSlflymJ9qg0bgOLFxeWptm2Bli3FZSsiIkPEcENEGVIogI4dxVpVY8aICQH37weqVAFGjQJevZK7QiIibQw3RJQlDg7A9OnA5ctAq1bA27fArFliQc7163mpiogMB8MNEWVL2bLAnj3Arl1A6dJi0r8uXYCGDYELF+SujoiI4YaIcujjj4F//gGmTQPs7ICjR4EaNYCBA8WyDkREcmG4IaIcs7UFxo4Frl4FPvsMUKmAxYvFLMcrV3JBTiKSB8MNEeVaiRLApk3AwYNiuPjTp0CfPsAHH4iZj4mI8hLDDRHpTNOmYm6cH34AHB2BM2dEwPnyS+DRI7mrIyJzwXBDRDqVLx8wbJiY5Tg4WGxbvVpcqpo3D3j9Wt76iMj0MdwQkV64uQEhIWK9qho1gNhYYPhwMRngqFHA3btyV0hEporhhoj0qm5d4NQpYPlyMXT8xQsxP07p0kD79mKUFefIISJdYrghIr2ztBQdjK9fB3bsAJo0ESOrtmwBGjQA/PxEK8+bN3JXSkSmgOGGiPKMpSXQpg0QHg5cvAj06iWGk58/L1YjL1ECmDBBTAxIRJRTBhFuFi9ejJIlS8LW1hZ16tTBqVOn0t135cqVqF+/PgoWLIiCBQsiICAgw/2JyDBVqSLmwvnvP2DGDNEX58kTYOpUwMsL+OILcTmLiCi7ZA83GzduxPDhwzFx4kScO3cOvr6+aN68OR4/fpzm/ocPH0bnzp1x6NAhnDhxAp6envjoo49w//79PK6ciHTBxQUYPRq4dQvYuBGoV0+sW7VuHVCnjuizs2GD2EZElBUKSZK3K1+dOnVQq1YtLFq0CACgUqng6emJQYMGYfTo0Zm+XqlUomDBgli0aBG6deuW6f6xsbFwdnZGTEwMnJyccl0/EenemTPAggXaoaZYMaB/f6B3b6BIEXnrI6K8l53Pb1lbbpKSknD27FkEBARotllYWCAgIAAnTpzI0jFev36Nt2/folChQvoqk4jyWM2awC+/APfuARMnAkWLAvfvi6UePD2Bnj1Fnx0iorTIGm6ePn0KpVIJV1dXre2urq6Ijo7O0jFGjRoFDw8PrYCUUmJiImJjY7VuRGQc3NyASZNEyPnlFzGqKjER+PlnwNcXaNwY2L6da1gRkTbZ+9zkxvfff48NGzZg27ZtsLW1TXOfGTNmwNnZWXPz9PTM4yqJKLdsbICuXYHTp4Fjx8QinZaWwOHDQNu2QJkywNy5wMuXcldKRIZA1nBTuHBhWFpa4tF7i848evQIbm5uGb52zpw5+P7773HgwAFUrVo13f3GjBmDmJgYzS0qKkontRNR3lMoAH9/sUjn7duiI3KhQsCdO8DIkWLE1YABYpVyIjJfsoYba2tr+Pn5ITw8XLNNpVIhPDwcdevWTfd1s2bNwtSpUxEaGoqaNWtmeA4bGxs4OTlp3YjI+Hl6iiHkUVHAihVApUpAfDywZAng4wO0aAHs2ycmCyQi8yL7Zanhw4dj5cqVWLNmDSIjI9GvXz/Ex8ejR48eAIBu3bphzJgxmv1nzpyJ8ePH4+eff0bJkiURHR2N6OhoxMXFyfUWiEhG9vZiBNWlS2JywDZtRAvP/v1Aq1ZAxYrA4sUA/4sgMh+yh5uOHTtizpw5mDBhAqpVq4aIiAiEhoZqOhnfu3cPD1NMV7p06VIkJSWhffv2cHd319zmzJkj11sgIgOgUIhlHXbsAG7cAIYOBZycgGvXgIEDxVDy4cPFfDpEZNpkn+cmr3GeGyLz8eoVsGaNmDPnxg2xTaEQrTtDhgCNGonHRGT4jGaeGyIifXJ0FK02V68Ce/YAH30kViBXL97p6wv89BOQkCB3pUSkSww3RGTyLCxE/5v9+4ErV4B+/URfnUuXRH8dT0/g22/FOldEZPwYbojIrPj4iBFV//0HzJ4tFul89kyMvCpZEujYEfjrL9HCQ0TGieGGiMxSwYJibpybN4EtW4CGDcVMx5s2ibl0atcG1q4FkpLkrpSIsovhhojMmpUV0K6dmO34/HmgRw8xI/KZM2JWZC8vYPJk4L25RonIgDHcEBH9X7VqYt2qqChg6lTA3R2IjhbrW5UoAQQHA+fOyV0lEWWGQ8GJiNKRlCQuWf34I3DyZPL2Dz4Ql7H8/MStVCkOKSfSt+x8fjPcEBFlwcmTYr6cTZuAd++0nytQAKhRIzns+PkB3t4MPES6xHCTAYYbIsqNBw+A3buBs2fF7dKltDsdOzsnBx711zJlxLB0Iso+hpsMMNwQkS4lJQGXL4u+OOrAc/EikJiYel8nJ6B6de0WnrJlGXiIsoLhJgMMN0Skb2/fAv/8I4KOOvRcuAC8eZN6XweH1IGnXDnA0jLv6yYyZAw3GWC4ISI5vH0LREYmt+6oA09aSz/kz68deGrUACpUYOAh88ZwkwGGGyIyFO/eiXWvUgaeiAjg9evU+9rbi6HqKVt4KlQQ8/QQmQOGmwww3BCRIVMqReBJ2Yfn/HkgPj71vnZ2YvHPlIGnYkUGHjJNDDcZYLghImOjVALXr2u38Jw/D8TFpd7X1jbtwJMvX97XTaRLDDcZYLghIlOgUgE3bmgHnnPngFevUu9rYwNUraodeCpVAqyt875uopxiuMkAww0RmSqVSiwEmvKS1rlzQExM6n2trUXAcXMDXFy0b4ULp95mb5/374coJYabDDDcEJE5UamAW7e0w87Zs8DLl9k7jq1t2qEno0Dk7Mw5fEh3GG4ywHBDROZOkoDbt8VcPE+fAs+eJd/ef/zsmRjGnhMWFkChQtkPRewfRGnJzuc3+9QTEZkZhQIoXVrcMiNJouNyWqEnvTD07Jl4jUolnn/6NHv1OTqmH3zSCkWFC4u5gYjUGG6IiChdCoUIG46OYvXzrEpMzF4YevoUePFChKlXr8Ttzp2sn8/LS0x8WK2a+Fq9OlC8OBcvNVe8LEVERAZBqRR9gbIShlI+l9bCpYBo1UkZdqpX59IWxox9bjLAcENEZDokSbT4XLwo5v6JiBBfr1wRYel9dnZiWHzKwFO5sthOho3hJgMMN0REpu/NG7FauzrsnD8v1vJKa2kLS0uxlIU67FSrJm6FCuVx0ZQhhpsMMNwQEZknpVJMfJgy8Jw/n36HZy+v1Je12I9HPgw3GWC4ISIiNUkCHjzQDjsREWKofFrYj0c+DDcZYLghIqLMvHyZ3MKT3X481aoBVaqwH4+uMdxkgOGGiIhyIqf9eNQtPezHkzsMNxlguCEiIl1RKsV6Xikva7Efj34w3GSA4YaIiPTp/X486paezPrxVKsGFC0qJkx0cEh9S7ndxsb8AhHDTQYYboiISA7qfjwpL2ul148nM5aWGYefnDxna2vYgYnhJgMMN0REZCjevBELmJ4/D1y6JAJQXFzq26tX4mtCgv5qsbDQXWAqUEC0SOkSF84kIiIyAra2gJ+fuGWFUgnEx6cdfDIKRRk9p+4QrVIBsbHills1awKnT+f+ODnFcENERGQkLC0BJydx0xWlUgQcXQYmR0fd1ZcTDDdERERmzNIyeeV3XZG7w4uFvKcnIiIiUyN3x2SGGyIiIjIpDDdERERkUhhuiIiIyKQw3BAREZFJYbghIiIik8JwQ0RERCaF4YaIiIhMCsMNERERmRSGGyIiIjIpDDdERERkUhhuiIiIyKQw3BAREZFJYbghIiIik2IldwF5Tfr/OuyxsbEyV0JERERZpf7cVn+OZ8Tsws2rV68AAJ6enjJXQkRERNn16tUrODs7Z7iPQspKBDIhKpUKDx48gKOjIxQKhdzlGKTY2Fh4enoiKioKTk5Ocpdj9vjzMCz8eRge/kwMi75+HpIk4dWrV/Dw8ICFRca9asyu5cbCwgLFixeXuwyj4OTkxP8oDAh/HoaFPw/Dw5+JYdHHzyOzFhs1digmIiIik8JwQ0RERCaF4YZSsbGxwcSJE2FjYyN3KQT+PAwNfx6Ghz8Tw2IIPw+z61BMREREpo0tN0RERGRSGG6IiIjIpDDcEBERkUlhuCEiIiKTwnBDGjNmzECtWrXg6OiIokWLIigoCNeuXZO7LALw/fffQ6FQYOjQoXKXYtbu37+PL774Ai4uLrCzs0OVKlVw5swZucsyS0qlEuPHj0epUqVgZ2cHb29vTJ06NUvrDlHu/fnnnwgMDISHhwcUCgW2b9+u9bwkSZgwYQLc3d1hZ2eHgIAA3LhxI8/qY7ghjSNHjmDAgAH4+++/ERYWhrdv3+Kjjz5CfHy83KWZtdOnT2P58uWoWrWq3KWYtRcvXsDf3x/58uXDvn37cOXKFcydOxcFCxaUuzSzNHPmTCxduhSLFi1CZGQkZs6ciVmzZmHhwoVyl2YW4uPj4evri8WLF6f5/KxZs7BgwQIsW7YMJ0+eRP78+dG8eXO8efMmT+rjUHBK15MnT1C0aFEcOXIEDRo0kLscsxQXF4caNWpgyZIlmDZtGqpVq4b58+fLXZZZGj16NI4fP46jR4/KXQoB+Pjjj+Hq6opVq1Zptn366aews7PD2rVrZazM/CgUCmzbtg1BQUEARKuNh4cHRowYgZEjRwIAYmJi4OrqipCQEHTq1EnvNbHlhtIVExMDAChUqJDMlZivAQMGoHXr1ggICJC7FLO3c+dO1KxZE5999hmKFi2K6tWrY+XKlXKXZbbq1auH8PBwXL9+HQBw4cIFHDt2DC1btpS5Mrp9+zaio6O1/t9ydnZGnTp1cOLEiTypwewWzqSsUalUGDp0KPz9/VG5cmW5yzFLGzZswLlz53D69Gm5SyEAt27dwtKlSzF8+HB8++23OH36NAYPHgxra2sEBwfLXZ7ZGT16NGJjY1GhQgVYWlpCqVTiu+++Q5cuXeQuzexFR0cDAFxdXbW2u7q6ap7TN4YbStOAAQNw+fJlHDt2TO5SzFJUVBSGDBmCsLAw2Nrayl0OQQT+mjVrYvr06QCA6tWr4/Lly1i2bBnDjQw2bdqEdevWYf369ahUqRIiIiIwdOhQeHh48OdBvCxFqQ0cOBC7d+/GoUOHULx4cbnLMUtnz57F48ePUaNGDVhZWcHKygpHjhzBggULYGVlBaVSKXeJZsfd3R0VK1bU2ubj44N79+7JVJF5+/rrrzF69Gh06tQJVapUQdeuXTFs2DDMmDFD7tLMnpubGwDg0aNHWtsfPXqkeU7fGG5IQ5IkDBw4ENu2bcMff/yBUqVKyV2S2WratCkuXbqEiIgIza1mzZro0qULIiIiYGlpKXeJZsff3z/V1AjXr1+Hl5eXTBWZt9evX8PCQvsjzNLSEiqVSqaKSK1UqVJwc3NDeHi4ZltsbCxOnjyJunXr5kkNvCxFGgMGDMD69euxY8cOODo6aq6NOjs7w87OTubqzIujo2Oqvk758+eHi4sL+0DJZNiwYahXrx6mT5+ODh064NSpU1ixYgVWrFghd2lmKTAwEN999x1KlCiBSpUq4fz58/jhhx/w5Zdfyl2aWYiLi8PNmzc1j2/fvo2IiAgUKlQIJUqUwNChQzFt2jSULVsWpUqVwvjx4+Hh4aEZUaV3EtH/AUjztnr1arlLI0mSGjZsKA0ZMkTuMszarl27pMqVK0s2NjZShQoVpBUrVshdktmKjY2VhgwZIpUoUUKytbWVSpcuLY0dO1ZKTEyUuzSzcOjQoTQ/L4KDgyVJkiSVSiWNHz9ecnV1lWxsbKSmTZtK165dy7P6OM8NERERmRT2uSEiIiKTwnBDREREJoXhhoiIiEwKww0RERGZFIYbIiIiMikMN0RERGRSGG6IiIjIpDDcEJFZUigU2L59u9xlEJEeMNwQUZ7r3r07FApFqluLFi3kLo2ITADXliIiWbRo0QKrV6/W2mZjYyNTNURkSthyQ0SysLGxgZubm9atYMGCAMQlo6VLl6Jly5aws7ND6dKl8fvvv2u9/tKlS2jSpAns7Ozg4uKCPn36IC4uTmufn3/+GZUqVYKNjQ3c3d0xcOBAreefPn2Ktm3bwt7eHmXLlsXOnTs1z7148QJdunRBkSJFYGdnh7Jly6YKY0RkmBhuiMggjR8/Hp9++ikuXLiALl26oFOnToiMjAQAxMfHo3nz5ihYsCBOnz6NzZs34+DBg1rhZenSpRgwYAD69OmDS5cuYefOnShTpozWOSZPnowOHTrg4sWLaNWqFbp06YLnz59rzn/lyhXs27cPkZGRWLp0KQoXLpx33wAiyrk8W6KTiOj/goODJUtLSyl//vxat++++06SJLFC/VdffaX1mjp16kj9+vWTJEmSVqxYIRUsWFCKi4vTPL9nzx7JwsJCio6OliRJkjw8PKSxY8emWwMAady4cZrHcXFxEgBp3759kiRJUmBgoNSjRw/dvGEiylPsc0NEsmjcuDGWLl2qta1QoUKa+3Xr1tV6rm7duoiIiAAAREZGwtfXF/nz59c87+/vD5VKhWvXrkGhUODBgwdo2rRphjVUrVpVcz9//vxwcnLC48ePAQD9+vXDp59+inPnzuGjjz5CUFAQ6tWrl6P3SkR5i+GGiGSRP3/+VJeJdMXOzi5L++XLl0/rsUKhgEqlAgC0bNkSd+/exd69exEWFoamTZtiwIABmDNnjs7rJSLdYp8bIjJIf//9d6rHPj4+AAAfHx9cuHAB8fHxmuePHz8OCwsLlC9fHo6OjihZsiTCw8NzVUORIkUQHByMtWvXYv78+VixYkWujkdEeYMtN0Qki8TERERHR2tts7Ky0nTa3bx5M2rWrIkPP/wQ69atw6lTp7Bq1SoAQJcuXTBx4kQEBwdj0qRJePLkCQYNGoSuXbvC1dUVADBp0iR89dVXKFq0KFq2bIlXr17h+PHjGDRoUJbqmzBhAvz8/FCpUiUkJiZi9+7dmnBFRIaN4YaIZBEaGgp3d3etbeXLl8fVq1cBiJFMGzZsQP/+/eHu7o7ffvsNFStWBADY29tj//79GDJkCGrVqgV7e3t8+umn+OGHHzTHCg4Oxps3bzBv3jyMHDkShQsXRvv27bNcn7W1NcaMGYM7d+7Azs4O9evXx4YNG3TwzolI3xSSJElyF0FElJJCocC2bdsQFBQkdylEZITY54aIiIhMCsMNERERmRT2uSEig8Or5USUG2y5ISIiIpPCcENEREQmheGGiIiITArDDREREZkUhhsiIiIyKQw3REREZFIYboiIiMikMNwQERGRSWG4ISIiIpPyP/c3fQgrfYRJAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Accessing the loss values\n",
        "train_loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "# Plotting the loss values\n",
        "epochs = range(1, len(train_loss) + 1)\n",
        "\n",
        "plt.plot(epochs, train_loss, 'b', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'r', label='Validation loss')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CHbxN6KWK1Ay",
        "outputId": "978a0a56-505c-4987-eb52-8467ebd7dd0c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 4852 images belonging to 5 classes.\n",
            "10/10 [==============================] - 17s 1s/step - loss: 0.6489 - accuracy: 0.8508\n",
            "Test loss: 0.6489183306694031\n",
            "Test accuracy: 0.8507831692695618\n"
          ]
        }
      ],
      "source": [
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "\n",
        "test_generator = test_datagen.flow_from_directory(\n",
        "        '/content/sample_data/miotcd/HW02-Practical/Dataset/test',\n",
        "        target_size=(128, 128),\n",
        "        batch_size=512,\n",
        "        class_mode='categorical')\n",
        "\n",
        "score = model.evaluate(test_generator)\n",
        "print('Test loss:', score[0])\n",
        "print('Test accuracy:', score[1])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iL-Gp126K7_6"
      },
      "source": [
        "# Best Fitting Model\n",
        "To perform hyperparameter tuning for our model using two different methods, specifically setting the learning rate using `grid search` and setting the learning rate using the `wandb` or `comet` libraries, we can follow the code snippets provided below."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kXkhUkzpL3Ho"
      },
      "source": [
        "**First, I'll demonstrate how to set the learning rate using grid search. For this purpose, we'll use the `GridSearchCV` class from the `sklearn` library. Here's an code:**"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "train_datagen = ImageDataGenerator(rescale=1./255)\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    '/content/drive/MyDrive/FCI-HW2/HW02-Practical/Dataset/train',\n",
        "    target_size=(128, 128),\n",
        "    batch_size=128,\n",
        "    class_mode='categorical'\n",
        ")\n",
        "\n",
        "validation_datagen = ImageDataGenerator(rescale=1./255)\n",
        "validation_generator = validation_datagen.flow_from_directory(\n",
        "    '/content/drive/MyDrive/FCI-HW2/HW02-Practical/Dataset/valid',\n",
        "    target_size=(128, 128),\n",
        "    batch_size=128,\n",
        "    class_mode='categorical'\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "b_osm6g11hhA",
        "outputId": "b4821465-3075-453a-940f-779293ac0a68"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 15000 images belonging to 5 classes.\n",
            "Found 5000 images belonging to 5 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sYogA3VgLzpa",
        "outputId": "1c3be73d-7e1e-4a07-f266-e044f2a98645"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-24-15afdd8771bb>:29: DeprecationWarning: KerasClassifier is deprecated, use Sci-Keras (https://github.com/adriangb/scikeras) instead. See https://www.adriangb.com/scikeras/stable/migration.html for help migrating.\n",
            "  model = KerasClassifier(build_fn=create_model, epochs=10, batch_size=128, verbose=0)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Best parameters:  {'learning_rate': 0.0001}\n",
            "Best score:  0.36674050490061444\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "# Define the model architecture\n",
        "def create_model(learning_rate):\n",
        "    model = Sequential()\n",
        "    model.add(Conv2D(64, (3, 3),activation='relu', input_shape=(128, 128, 3)))\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(Conv2D(128, (3, 3),activation='relu'))\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(Conv2D(256, (3, 3),activation='relu'))\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(Flatten())\n",
        "    model.add(Dense(512, activation='relu'))\n",
        "    model.add(Dropout(0.5))\n",
        "    model.add(Dense(256, activation='relu'))\n",
        "    model.add(Dropout(0.5))\n",
        "    model.add(Dense(5, activation='softmax'))\n",
        "\n",
        "    optimizer = Adam(learning_rate=learning_rate)\n",
        "    model.compile(loss='categorical_crossentropy', optimizer=optimizer, metrics=['accuracy'])\n",
        "    return model\n",
        "\n",
        "# Create the Keras classifier for use in GridSearchCV\n",
        "model = KerasClassifier(build_fn=create_model, epochs=10, batch_size=128, verbose=0)\n",
        "\n",
        "# Define the grid search parameters\n",
        "learning_rate = [0.0001, 0.001, 0.01, 0.1]\n",
        "\n",
        "param_grid = dict(learning_rate=learning_rate)\n",
        "\n",
        "# Load the data from the generator\n",
        "X_train, y_train = next(train_generator)\n",
        "X_val, y_val = next(validation_generator)\n",
        "\n",
        "# Create the grid search\n",
        "grid = GridSearchCV(estimator=model, param_grid=param_grid, cv=3)\n",
        "grid_result = grid.fit(X_train, y_train, validation_data=(X_val, y_val))\n",
        "\n",
        "# Print the best parameters and score\n",
        "print(\"Best parameters: \", grid_result.best_params_)\n",
        "print(\"Best score: \", grid_result.best_score_)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hINMAL2P28xB"
      },
      "source": [
        "**now we plot the best model loss**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "Ezax9lvL28Ou",
        "outputId": "7d8efa19-ed0d-46f2-cd99-e6ba121600e5"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Get the best parameters and corresponding model\n",
        "best_params = grid_result.best_params_\n",
        "best_model = grid_result.best_estimator_.model\n",
        "\n",
        "# Train the model with the best parameters\n",
        "history = best_model.fit(X_train, y_train, validation_data=(X_val, y_val), epochs=10, batch_size=128, verbose=0)\n",
        "\n",
        "# Plot the loss on the training and validation sets\n",
        "plt.plot(history.history['loss'], label='Training 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()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "blvgOJcONSfF"
      },
      "source": [
        "**Now, let's move on to setting the learning rate using the `wandb` or `comet` libraries. These libraries allow us to track and log experiments, including hyperparameters. Here's an code snippets using `wandb`:**"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import wandb\n",
        "\n",
        "wandb.login()\n",
        "\n",
        "\n",
        "# Define sweep config\n",
        "sweep_configuration = {\n",
        "    'method': 'grid',\n",
        "    'name': 'sweep_grid',\n",
        "    'metric': {'goal': 'maximize', 'name': 'val_accuracy'},\n",
        "    'parameters': {\n",
        "        'learning_rate': {'values': [1e-3, 1e-4, 1e-5]}\n",
        "    }\n",
        "}\n",
        "\n",
        "# Default values for hyper-parameters we're going to sweep over\n",
        "configs = {\n",
        "       'learning_rate' : 0.001\n",
        "   }\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RZPZmJvb4b-q",
        "outputId": "e2c17948-1c68-4838-be04-4d3fc0c259c6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mmr-srezaei\u001b[0m. Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Initialize sweep by passing in the config\n",
        "sweep_id = wandb.sweep(sweep_configuration, project=\"miotcd\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R7GW_utdBLhy",
        "outputId": "8a25660a-b4f3-4971-ac31-f9a874da7522"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Network error (ReadTimeout), entering retry loop.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Create sweep with ID: 7zgqhsyq\n",
            "Sweep URL: https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "u9cD7xUmNqKI"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from wandb.keras import WandbCallback\n",
        "\n",
        "\n",
        "def train(config = None):\n",
        "    with wandb.init(config=config):\n",
        "        config = wandb.config\n",
        "\n",
        "        model = Sequential()\n",
        "        model.add(Conv2D(64, (3, 3),activation='relu', input_shape=(128, 128, 3)))\n",
        "        model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "        model.add(Conv2D(128, (3, 3),activation='relu'))\n",
        "        model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "        model.add(Conv2D(256, (3, 3),activation='relu'))\n",
        "        model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "\n",
        "        model.add(Flatten())\n",
        "        model.add(Dense(512, activation='relu'))\n",
        "        model.add(Dropout(0.5))\n",
        "        model.add(Dense(256, activation='relu'))\n",
        "        model.add(Dropout(0.5))\n",
        "        model.add(Dense(5, activation='softmax'))\n",
        "\n",
        "\n",
        "        # Set the learning rate using wandb\n",
        "        optimizer = Adam(learning_rate=config.learning_rate)\n",
        "        model.compile(loss='categorical_crossentropy', optimizer=optimizer, metrics=['accuracy'])\n",
        "\n",
        "        # Train the model and log the metrics\n",
        "        model.fit(train_generator,\n",
        "              steps_per_epoch=train_generator.n // train_generator.batch_size,\n",
        "              epochs=10,\n",
        "              validation_data=validation_generator,\n",
        "              validation_steps=validation_generator.n // validation_generator.batch_size,\n",
        "              callbacks=[WandbCallback()])\n",
        "        wandb.log({'train_loss': model.history.history['loss'], 'val_loss': model.history.history['val_loss']})"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "wandb.agent(sweep_id, train, count=5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000,
          "referenced_widgets": [
            "16e7eaf142c24de298ef72bb74da9521",
            "ca6439a995574a83bd4131583127f6e8",
            "d8a64efba9694880b23931ea7ac3dd7b",
            "3af53fc5aa4a470abe230aef46cdf584",
            "1adcce1812a84dffa6e510f1d556644d",
            "ce71f5b3dd164f6683cc62cc6931688b",
            "69604f7275b34bdca459ccf439217697",
            "ff576fc9199443f491b3b9f3e20ec943"
          ]
        },
        "id": "MQqCDYdmCqtD",
        "outputId": "89a0991f-e0ae-4882-8292-39b8d3b65d6f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: jbhenuzr with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlearning_rate: 0.001\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230521_204112-jbhenuzr</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/miotcd/runs/jbhenuzr' target=\"_blank\">zesty-sweep-1</a></strong> to <a href='https://wandb.ai/mr-srezaei/miotcd' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/miotcd' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/miotcd/runs/jbhenuzr' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/runs/jbhenuzr</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m The save_model argument by default saves the model in the HDF5 format that cannot save custom objects like subclassed models and custom layers. This behavior will be deprecated in a future release in favor of the SavedModel format. Meanwhile, the HDF5 model is saved as W&B files and the SavedModel as W&B Artifacts.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 1.1485 - accuracy: 0.5381"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_204112-jbhenuzr/files/model-best)... Done. 1.4s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 79s 610ms/step - loss: 1.1485 - accuracy: 0.5381 - val_loss: 0.7554 - val_accuracy: 0.7183\n",
            "Epoch 2/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.6320 - accuracy: 0.7715"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_204112-jbhenuzr/files/model-best)... Done. 1.4s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 71s 605ms/step - loss: 0.6320 - accuracy: 0.7715 - val_loss: 0.6022 - val_accuracy: 0.7881\n",
            "Epoch 3/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.4469 - accuracy: 0.8448"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_204112-jbhenuzr/files/model-best)... Done. 1.8s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 60s 516ms/step - loss: 0.4469 - accuracy: 0.8448 - val_loss: 0.5489 - val_accuracy: 0.8259\n",
            "Epoch 4/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.3106 - accuracy: 0.8959"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_204112-jbhenuzr/files/model-best)... Done. 1.5s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 68s 585ms/step - loss: 0.3106 - accuracy: 0.8959 - val_loss: 0.4516 - val_accuracy: 0.8534\n",
            "Epoch 5/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.2243 - accuracy: 0.9225"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_204112-jbhenuzr/files/model-best)... Done. 1.5s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 69s 587ms/step - loss: 0.2243 - accuracy: 0.9225 - val_loss: 0.4276 - val_accuracy: 0.8744\n",
            "Epoch 6/10\n",
            "117/117 [==============================] - 55s 467ms/step - loss: 0.1726 - accuracy: 0.9406 - val_loss: 0.4545 - val_accuracy: 0.8592\n",
            "Epoch 7/10\n",
            "117/117 [==============================] - 62s 532ms/step - loss: 0.1297 - accuracy: 0.9584 - val_loss: 0.4372 - val_accuracy: 0.8780\n",
            "Epoch 8/10\n",
            "117/117 [==============================] - 53s 450ms/step - loss: 0.1056 - accuracy: 0.9664 - val_loss: 0.4651 - val_accuracy: 0.8722\n",
            "Epoch 9/10\n",
            "117/117 [==============================] - 61s 522ms/step - loss: 0.0834 - accuracy: 0.9727 - val_loss: 0.5455 - val_accuracy: 0.8776\n",
            "Epoch 10/10\n",
            "117/117 [==============================] - 52s 446ms/step - loss: 0.0667 - accuracy: 0.9788 - val_loss: 0.5951 - val_accuracy: 0.8710\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▁▅▆▇▇▇████</td></tr><tr><td>epoch</td><td>▁▂▃▃▄▅▆▆▇█</td></tr><tr><td>loss</td><td>█▅▃▃▂▂▁▁▁▁</td></tr><tr><td>val_accuracy</td><td>▁▄▆▇█▇████</td></tr><tr><td>val_loss</td><td>█▅▄▂▁▂▁▂▄▅</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.97878</td></tr><tr><td>best_epoch</td><td>4</td></tr><tr><td>best_val_loss</td><td>0.42758</td></tr><tr><td>epoch</td><td>9</td></tr><tr><td>loss</td><td>0.06673</td></tr><tr><td>val_accuracy</td><td>0.87099</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">zesty-sweep-1</strong> at: <a href='https://wandb.ai/mr-srezaei/miotcd/runs/jbhenuzr' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/runs/jbhenuzr</a><br/>Synced 5 W&B file(s), 1 media file(s), 25 artifact file(s) and 1 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230521_204112-jbhenuzr/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: l7jfdutt with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlearning_rate: 0.0001\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230521_205318-l7jfdutt</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/miotcd/runs/l7jfdutt' target=\"_blank\">pretty-sweep-2</a></strong> to <a href='https://wandb.ai/mr-srezaei/miotcd' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/miotcd' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/miotcd/runs/l7jfdutt' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/runs/l7jfdutt</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 1.2488 - accuracy: 0.4884"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 1.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 77s 645ms/step - loss: 1.2488 - accuracy: 0.4884 - val_loss: 0.8769 - val_accuracy: 0.6937\n",
            "Epoch 2/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.8659 - accuracy: 0.6816"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 6.9s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 73s 627ms/step - loss: 0.8659 - accuracy: 0.6816 - val_loss: 0.7876 - val_accuracy: 0.7165\n",
            "Epoch 3/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.6937 - accuracy: 0.7523"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 2.1s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 61s 518ms/step - loss: 0.6937 - accuracy: 0.7523 - val_loss: 0.6992 - val_accuracy: 0.7610\n",
            "Epoch 4/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.5837 - accuracy: 0.7915"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 1.8s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 67s 572ms/step - loss: 0.5837 - accuracy: 0.7915 - val_loss: 0.6243 - val_accuracy: 0.7780\n",
            "Epoch 5/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.5034 - accuracy: 0.8243"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 1.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 67s 575ms/step - loss: 0.5034 - accuracy: 0.8243 - val_loss: 0.5487 - val_accuracy: 0.8151\n",
            "Epoch 6/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.4339 - accuracy: 0.8542"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 3.0s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 72s 617ms/step - loss: 0.4339 - accuracy: 0.8542 - val_loss: 0.5038 - val_accuracy: 0.8323\n",
            "Epoch 7/10\n",
            "117/117 [==============================] - 52s 447ms/step - loss: 0.3767 - accuracy: 0.8730 - val_loss: 0.5365 - val_accuracy: 0.8285\n",
            "Epoch 8/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.3216 - accuracy: 0.8912"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 1.8s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 68s 588ms/step - loss: 0.3216 - accuracy: 0.8912 - val_loss: 0.4634 - val_accuracy: 0.8476\n",
            "Epoch 9/10\n",
            "117/117 [==============================] - 63s 538ms/step - loss: 0.2861 - accuracy: 0.9063 - val_loss: 0.4753 - val_accuracy: 0.8385\n",
            "Epoch 10/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.2563 - accuracy: 0.9160"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_205318-l7jfdutt/files/model-best)... Done. 4.6s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 72s 617ms/step - loss: 0.2563 - accuracy: 0.9160 - val_loss: 0.4443 - val_accuracy: 0.8564\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▁▄▅▆▆▇▇███</td></tr><tr><td>epoch</td><td>▁▂▃▃▄▅▆▆▇█</td></tr><tr><td>loss</td><td>█▅▄▃▃▂▂▁▁▁</td></tr><tr><td>val_accuracy</td><td>▁▂▄▅▆▇▇█▇█</td></tr><tr><td>val_loss</td><td>█▇▅▄▃▂▂▁▂▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.91601</td></tr><tr><td>best_epoch</td><td>9</td></tr><tr><td>best_val_loss</td><td>0.44431</td></tr><tr><td>epoch</td><td>9</td></tr><tr><td>loss</td><td>0.25628</td></tr><tr><td>val_accuracy</td><td>0.85637</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">pretty-sweep-2</strong> at: <a href='https://wandb.ai/mr-srezaei/miotcd/runs/l7jfdutt' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/runs/l7jfdutt</a><br/>Synced 5 W&B file(s), 1 media file(s), 40 artifact file(s) and 1 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230521_205318-l7jfdutt/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: 6qpf5tq9 with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlearning_rate: 1e-05\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230521_210558-6qpf5tq9</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/miotcd/runs/6qpf5tq9' target=\"_blank\">swift-sweep-3</a></strong> to <a href='https://wandb.ai/mr-srezaei/miotcd' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/miotcd' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/sweeps/7zgqhsyq</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/miotcd/runs/6qpf5tq9' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/runs/6qpf5tq9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 1.5452 - accuracy: 0.3141"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 3.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 72s 588ms/step - loss: 1.5452 - accuracy: 0.3141 - val_loss: 1.3889 - val_accuracy: 0.5853\n",
            "Epoch 2/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 1.2785 - accuracy: 0.5172"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 1.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 74s 630ms/step - loss: 1.2785 - accuracy: 0.5172 - val_loss: 1.0914 - val_accuracy: 0.6102\n",
            "Epoch 3/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 1.0983 - accuracy: 0.5887"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 1.9s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 61s 520ms/step - loss: 1.0983 - accuracy: 0.5887 - val_loss: 0.9626 - val_accuracy: 0.6675\n",
            "Epoch 4/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 1.0072 - accuracy: 0.6296"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 2.3s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 63s 541ms/step - loss: 1.0072 - accuracy: 0.6296 - val_loss: 0.9125 - val_accuracy: 0.6707\n",
            "Epoch 5/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.9463 - accuracy: 0.6497"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 4.4s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 79s 679ms/step - loss: 0.9463 - accuracy: 0.6497 - val_loss: 0.8700 - val_accuracy: 0.6921\n",
            "Epoch 6/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.8953 - accuracy: 0.6733"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 2.0s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 73s 625ms/step - loss: 0.8953 - accuracy: 0.6733 - val_loss: 0.8355 - val_accuracy: 0.6969\n",
            "Epoch 7/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.8615 - accuracy: 0.6823"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 1.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 68s 582ms/step - loss: 0.8615 - accuracy: 0.6823 - val_loss: 0.8085 - val_accuracy: 0.7131\n",
            "Epoch 8/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.8312 - accuracy: 0.6972"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 5.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 76s 650ms/step - loss: 0.8312 - accuracy: 0.6972 - val_loss: 0.7864 - val_accuracy: 0.7196\n",
            "Epoch 9/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.8053 - accuracy: 0.7096"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 6.8s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 78s 667ms/step - loss: 0.8053 - accuracy: 0.7096 - val_loss: 0.7709 - val_accuracy: 0.7306\n",
            "Epoch 10/10\n",
            "117/117 [==============================] - ETA: 0s - loss: 0.7830 - accuracy: 0.7137"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Adding directory to artifact (/content/wandb/run-20230521_210558-6qpf5tq9/files/model-best)... Done. 1.7s\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r117/117 [==============================] - 62s 530ms/step - loss: 0.7830 - accuracy: 0.7137 - val_loss: 0.7590 - val_accuracy: 0.7344\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "VBox(children=(Label(value='3300.153 MB of 3300.167 MB uploaded (0.205 MB deduped)\\r'), FloatProgress(value=0.…"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "16e7eaf142c24de298ef72bb74da9521"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▁▅▆▇▇▇▇███</td></tr><tr><td>epoch</td><td>▁▂▃▃▄▅▆▆▇█</td></tr><tr><td>loss</td><td>█▆▄▃▂▂▂▁▁▁</td></tr><tr><td>val_accuracy</td><td>▁▂▅▅▆▆▇▇██</td></tr><tr><td>val_loss</td><td>█▅▃▃▂▂▂▁▁▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.71366</td></tr><tr><td>best_epoch</td><td>9</td></tr><tr><td>best_val_loss</td><td>0.75898</td></tr><tr><td>epoch</td><td>9</td></tr><tr><td>loss</td><td>0.78304</td></tr><tr><td>val_accuracy</td><td>0.73438</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">swift-sweep-3</strong> at: <a href='https://wandb.ai/mr-srezaei/miotcd/runs/6qpf5tq9' target=\"_blank\">https://wandb.ai/mr-srezaei/miotcd/runs/6qpf5tq9</a><br/>Synced 5 W&B file(s), 1 media file(s), 50 artifact file(s) and 1 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230521_210558-6qpf5tq9/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Sweep Agent: Waiting for job.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Sweep Agent: Exiting.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Evaluating Files\n",
        "in the code snippets below we save our evaluating scores into csv file"
      ],
      "metadata": {
        "id": "A2TkENIEpvmF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install wandb"
      ],
      "metadata": {
        "id": "ZRXxqU-rp-Aq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import wandb\n",
        "import tensorflow as tf\n",
        "wandb.login()\n",
        "\n",
        "run = wandb.init()\n",
        "artifact = run.use_artifact('mr-srezaei/miotcd/model-zesty-sweep-1:v4', type='model')\n",
        "artifact_dir = artifact.download()\n",
        "\n",
        "best_model = tf.keras.models.load_model(artifact_dir)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "ZMUPIfjx8nvc",
        "outputId": "b4b589f2-6e94-4fa7-cbfa-b70293323125"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m Calling wandb.login() after wandb.init() has no effect.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Finishing last run (ID:y0kkc9qc) before initializing another..."
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">visionary-glade-2</strong> at: <a href='https://wandb.ai/mr-srezaei/uncategorized/runs/y0kkc9qc' target=\"_blank\">https://wandb.ai/mr-srezaei/uncategorized/runs/y0kkc9qc</a><br/>Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230523_172915-y0kkc9qc/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Successfully finished last run (ID:y0kkc9qc). Initializing new run:<br/>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230523_184413-lesfzdnf</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/uncategorized/runs/lesfzdnf' target=\"_blank\">vital-plasma-2</a></strong> to <a href='https://wandb.ai/mr-srezaei/uncategorized' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/uncategorized' target=\"_blank\">https://wandb.ai/mr-srezaei/uncategorized</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/uncategorized/runs/lesfzdnf' target=\"_blank\">https://wandb.ai/mr-srezaei/uncategorized/runs/lesfzdnf</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Downloading large artifact model-zesty-sweep-1:v4, 300.03MB. 5 files... \n",
            "\u001b[34m\u001b[1mwandb\u001b[0m:   5 of 5 files downloaded.  \n",
            "Done. 0:0:1.8\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from os import listdir\n",
        "from os.path import isfile, join, splitext\n",
        "from tqdm import tqdm\n",
        "import numpy as np\n",
        "import csv\n",
        "import pandas as pd\n",
        "\n",
        "# Define your classes\n",
        "classes = ['articulated_truck', 'bus', 'car', 'work_van', 'background']\n",
        "\n",
        "def load_and_preprocess_image(image_path):\n",
        "    # Read image from file\n",
        "    image = tf.io.read_file(image_path)\n",
        "    \n",
        "    # Decode image and convert to desired format\n",
        "    image = tf.image.decode_image(image, channels=3)\n",
        "    \n",
        "    # Resize the image to (128, 128)\n",
        "    image = tf.image.resize(image, [128, 128])\n",
        "    \n",
        "    # Rescale the pixel values to the range [0, 1]\n",
        "    image = image / 255.0\n",
        "    \n",
        "    return image\n",
        "\n",
        "def classify_image(path_to_image):\n",
        "\n",
        "    # Load the image and preprocess it if needed\n",
        "    image = load_and_preprocess_image(path_to_image)\n",
        "\n",
        "    # Make predictions using your trained model\n",
        "    predictions = best_model.predict(np.array([image]))  # Assuming 'model' is your trained model\n",
        "\n",
        "    # Get the predicted label\n",
        "    label_index = np.argmax(predictions[0])\n",
        "    label = classes[label_index]\n",
        "\n",
        "    return label\n",
        "\n",
        "def parse_dataset(path_to_dataset):\n",
        "    '''\n",
        "    Parse every image contained in 'path_to_dataset' (a path to the training\n",
        "    or testing set), classify each image and save in a csv file the resulting\n",
        "    assignment\n",
        "\n",
        "    dataset_result: dict structure returned by the function.  It contains the\n",
        "            label of each image\n",
        "    '''\n",
        "    llist = listdir(path_to_dataset)\n",
        "    dataset_result = {}\n",
        "\n",
        "    for name in tqdm(llist):\n",
        "        dn = join(path_to_dataset, name)\n",
        "        if isfile(dn):\n",
        "            label = classify_image(dn)\n",
        "            file_nb, file_ext = splitext(name)\n",
        "            dataset_result[file_nb] = label\n",
        "\n",
        "        else:\n",
        "            file_list = listdir(dn)\n",
        "            for file_name in file_list:\n",
        "                file_name_with_path = join(dn, file_name)\n",
        "                label = classify_image(file_name_with_path)\n",
        "                file_nb, file_ext = splitext(file_name)\n",
        "                if file_nb in dataset_result.keys():\n",
        "                    print('error! ', file_nb, dataset_result[file_nb], ' vs ', file_name_with_path)\n",
        "                dataset_result[file_nb] = label\n",
        "\n",
        "    return dataset_result\n",
        "\n",
        "def save_classification_result(dataset_result, output_csv_file_name):\n",
        "    '''\n",
        "    save the dataset_result (a dict structure containing the class of every image)\n",
        "    into a valid csv file.\n",
        "    '''\n",
        "\n",
        "    with open(output_csv_file_name, 'w', newline='') as csvfile:\n",
        "        fieldnames = ['file_name', 'class_label']\n",
        "        writer = csv.DictWriter(csvfile, fieldnames=fieldnames)\n",
        "        writer.writeheader()\n",
        "\n",
        "        for file_name, class_label in dataset_result.items():\n",
        "            file_name_without_zeros = file_name.lstrip('0')\n",
        "            writer.writerow({'file_name': file_name_without_zeros, 'class_label': class_label})\n",
        "\n",
        "    csvfile.close()\n",
        "\n",
        "if __name__ == '__main__':\n",
        "    # Specify the path to the test dataset and the output CSV file name\n",
        "    path_to_test_dataset = '/content/sample_data/miotcd/HW02-Practical/Dataset/test'  # Replace with your actual path\n",
        "    output_csv_file_name = 'your_results_test.csv'  # Replace with your desired output file name\n",
        "\n",
        "    dataset_result = parse_dataset(path_to_test_dataset)\n",
        "    save_classification_result(dataset_result, output_csv_file_name)"
      ],
      "metadata": {
        "id": "sokt7-DTF4XU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import csv\n",
        "from sklearn.metrics import confusion_matrix, cohen_kappa_score, precision_score, recall_score, f1_score\n",
        "import numpy as np\n",
        "\n",
        "def compute_metric(result, gt, classes):\n",
        "    y_true = []\n",
        "    y_pred = []\n",
        "\n",
        "\n",
        "    for img in gt.keys():\n",
        "        if img not in result:\n",
        "            raise ValueError(f\"Image '{img}' missing in the result data.\")\n",
        "        \n",
        "        y_true.append(gt[img])\n",
        "        y_pred.append(result[img])\n",
        "\n",
        "    cm = confusion_matrix(y_true, y_pred, labels=classes)\n",
        "    cks = cohen_kappa_score(y_true, y_pred, labels=classes)\n",
        "    ps = precision_score(y_true, y_pred, average=None)\n",
        "    rs = recall_score(y_true, y_pred, average=None)\n",
        "    fs = f1_score(y_true, y_pred, average=None)\n",
        "\n",
        "    metrics = {}\n",
        "    metrics['confusion matrix'] = cm\n",
        "    metrics['cohen kappa score'] = cks\n",
        "    metrics['precision score'] = ps\n",
        "    metrics['recall score'] = rs\n",
        "    metrics['f1 score'] = fs\n",
        "    metrics['mean recall'] = np.mean(rs)\n",
        "    metrics['mean precision'] = np.mean(ps)\n",
        "    metrics['accuracy'] = np.diagonal(cm).sum() / float(len(y_true))\n",
        "    return metrics\n",
        "\n",
        "def csv_evaluation(gt_file, res_file, classes):\n",
        "    gt = {}\n",
        "    with open(gt_file, 'r') as csv_file:\n",
        "        reader = csv.reader(csv_file, delimiter=',')\n",
        "        for row in reader:\n",
        "            gt[row[0]] = row[1]\n",
        "\n",
        "\n",
        "    results = {}\n",
        "    with open(res_file, 'r') as csv_file:\n",
        "        reader = csv.reader(csv_file, delimiter=',')\n",
        "        for row in reader:\n",
        "            results[row[0]] = row[1]\n",
        "\n",
        "\n",
        "    metrics = compute_metric(results, gt, classes)\n",
        "    return metrics\n",
        "\n",
        "if __name__ == '__main__':\n",
        "    gt_file = '/content/drive/MyDrive/Colab Notebooks/HW-2/gt_test.csv'\n",
        "    res_file = '/content/drive/MyDrive/Colab Notebooks/HW-2/your_results_test.csv'\n",
        "    classes = ['articulated_truck', 'bus', 'car', 'work_van', 'background']  # Provide the list of classes\n",
        "\n",
        "    print('Computing score between', gt_file, 'and', res_file, '\\n')\n",
        "    metrics = csv_evaluation(gt_file, res_file, classes)\n",
        "\n",
        "    for key in metrics.keys():\n",
        "        print(key + ':')\n",
        "        print(metrics[key])\n",
        "        print('\\n')\n"
      ],
      "metadata": {
        "id": "D7F-lmnYF4JB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6f55b5da-24d9-4f2c-c3a3-530bb57413c0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Computing score between /content/drive/MyDrive/Colab Notebooks/HW-2/gt_test.csv and /content/drive/MyDrive/Colab Notebooks/HW-2/your_results_test.csv \n",
            "\n",
            "confusion matrix:\n",
            "[[875  46  28  18  33]\n",
            " [ 40   9 913   9  29]\n",
            " [ 53  20  33 738 156]\n",
            " [ 51  61  36  63 789]\n",
            " [ 10 977   2   7   4]]\n",
            "\n",
            "\n",
            "cohen kappa score:\n",
            "-0.0040000000000000036\n",
            "\n",
            "\n",
            "precision score:\n",
            "[0.85034014 0.00395648 0.00808625 0.0326087  0.0754491 ]\n",
            "\n",
            "\n",
            "recall score:\n",
            "[0.875 0.004 0.009 0.033 0.063]\n",
            "\n",
            "\n",
            "f1 score:\n",
            "[0.86249384 0.00397812 0.00851869 0.03280318 0.06866485]\n",
            "\n",
            "\n",
            "mean recall:\n",
            "0.1968\n",
            "\n",
            "\n",
            "mean precision:\n",
            "0.19408813312124037\n",
            "\n",
            "\n",
            "accuracy:\n",
            "0.1968\n",
            "\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# <font color='red'># Part 2</font>"
      ],
      "metadata": {
        "id": "oH6PtGBL32He"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Tranfer learning Optimize using Bayesian"
      ],
      "metadata": {
        "id": "A4WLFMQb-nO1"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "in this phase we use `resnet18` to create a <font color = 'blue'>transfer learning</font> model and train it with our `MIO-TCD` dataset.then tune the `learning rate` hyper parameter using `wandb` library by `random search` method."
      ],
      "metadata": {
        "id": "PXtoKlYw4FrS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "from torch.utils.data import DataLoader\n",
        "from torchvision import datasets, models, transforms\n",
        "import time"
      ],
      "metadata": {
        "id": "Ewh0Nwj6360p"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import wandb\n",
        "\n",
        "wandb.login()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "id": "OEQKmiEqyyEs",
        "outputId": "308d26d5-26b7-48a9-d3f9-d7536e04622b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "\n",
              "        window._wandbApiKey = new Promise((resolve, reject) => {\n",
              "            function loadScript(url) {\n",
              "            return new Promise(function(resolve, reject) {\n",
              "                let newScript = document.createElement(\"script\");\n",
              "                newScript.onerror = reject;\n",
              "                newScript.onload = resolve;\n",
              "                document.body.appendChild(newScript);\n",
              "                newScript.src = url;\n",
              "            });\n",
              "            }\n",
              "            loadScript(\"https://cdn.jsdelivr.net/npm/postmate/build/postmate.min.js\").then(() => {\n",
              "            const iframe = document.createElement('iframe')\n",
              "            iframe.style.cssText = \"width:0;height:0;border:none\"\n",
              "            document.body.appendChild(iframe)\n",
              "            const handshake = new Postmate({\n",
              "                container: iframe,\n",
              "                url: 'https://wandb.ai/authorize'\n",
              "            });\n",
              "            const timeout = setTimeout(() => reject(\"Couldn't auto authenticate\"), 5000)\n",
              "            handshake.then(function(child) {\n",
              "                child.on('authorize', data => {\n",
              "                    clearTimeout(timeout)\n",
              "                    resolve(data)\n",
              "                });\n",
              "            });\n",
              "            })\n",
              "        });\n",
              "    "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Logging into wandb.ai. (Learn how to deploy a W&B server locally: https://wandb.me/wandb-server)\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: You can find your API key in your browser here: https://wandb.ai/authorize\n",
            "wandb: Paste an API key from your profile and hit enter, or press ctrl+c to quit:"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            " ··········\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Appending key for api.wandb.ai to your netrc file: /root/.netrc\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sweep_configuration = {\n",
        "    'method': 'bayes',\n",
        "    'name': 'sweep',\n",
        "    'metric': {\n",
        "        'goal': 'minimize', \n",
        "        'name': 'validation_loss'\n",
        "        },\n",
        "    'parameters': {\n",
        "        'lr': {'values': [1e-3, 1e-4, 1e-5]}\n",
        "     }\n",
        "}\n",
        "\n",
        "sweep_id = wandb.sweep(sweep_configuration, project='transfer')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QTcdhR2sy1lz",
        "outputId": "28aec0cd-a551-43d5-b758-c95584b90f92"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Create sweep with ID: yoabtqx9\n",
            "Sweep URL: https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Define the data transformations for training and validation sets\n",
        "data_transforms = {\n",
        "    'train': transforms.Compose([\n",
        "        transforms.RandomResizedCrop(128), # Randomly crop the image to size 128x128\n",
        "        transforms.RandomHorizontalFlip(), # Randomly flip the image horizontally\n",
        "        transforms.ToTensor(),  # Convert the image to a tensor\n",
        "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
        "    ]),\n",
        "    'val': transforms.Compose([\n",
        "        transforms.Resize((128, 128)),  # Resize the image to 128x128\n",
        "        transforms.ToTensor(), # Convert the image to a tensor\n",
        "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]) # Normalize the image with mean and standard deviation\n",
        "    ]),\n",
        "    'test': transforms.Compose([\n",
        "        transforms.Resize((128, 128)),  # Resize the image to 128x128\n",
        "        transforms.ToTensor(), # Convert the image to a tensor\n",
        "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]) # Normalize the image with mean and standard deviation\n",
        "    ]),\n",
        "}"
      ],
      "metadata": {
        "id": "EuEHAYC94rQE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Set the directory paths for the dataset\n",
        "data_dir = '/content/sample_data/miotcd/HW02-Practical/Dataset'\n",
        "train_dir = data_dir + '/train'\n",
        "val_dir = data_dir + '/valid'\n",
        "test_dir = data_dir + '/test'"
      ],
      "metadata": {
        "id": "f6gcpTm_5ADY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the training and validation datasets using ImageFolder\n",
        "train_data = datasets.ImageFolder(train_dir, transform=data_transforms['train'])\n",
        "val_data = datasets.ImageFolder(val_dir, transform=data_transforms['val'])\n",
        "test_data= datasets.ImageFolder(test_dir, transform=data_transforms['test'])\n",
        "\n",
        "# Create data loaders for training and validation\n",
        "train_loader = DataLoader(train_data, batch_size=256, shuffle=True, num_workers=2)\n",
        "val_loader = DataLoader(val_data, batch_size=128, num_workers=2)\n",
        "test_loader = DataLoader(test_data, batch_size=128, num_workers=2)\n",
        "\n",
        "# Define the device for training (GPU if available, otherwise CPU)\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
      ],
      "metadata": {
        "id": "Io1B-jQV5FrK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print('Train dataset size:', len(train_data))\n",
        "print('valid dataset size:', len(val_data))\n",
        "print('test dataset size:', len(test_data))\n",
        "class_names = train_data.classes\n",
        "print('Class names:', class_names)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oRv23tzv-e1n",
        "outputId": "c1a3cfea-fde4-43a6-a13e-f7780c531d2f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train dataset size: 15000\n",
            "valid dataset size: 5000\n",
            "test dataset size: 5000\n",
            "Class names: ['articulated_truck', 'background', 'bus', 'car', 'work_van']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the pre-trained ResNet-18 model\n",
        "model = models.resnet18(pretrained=True)\n",
        "\n",
        "# Freeze the parameters of the pre-trained model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Get the number of input features for the last fully connected layer\n",
        "num_features = model.fc.in_features\n",
        "\n",
        "# Replace the last fully connected layer with a new one (output size 5)\n",
        "model.fc = nn.Linear(num_features, 5)\n",
        "\n",
        "# Transfer the model to the device\n",
        "model = model.to(device)\n",
        "\n",
        "# Define the loss function\n",
        "criterion = nn.CrossEntropyLoss()"
      ],
      "metadata": {
        "id": "63eDxFss5RLh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "029b974a-2f69-43e9-edff-009b95c26115"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n",
            "  warnings.warn(\n",
            "/usr/local/lib/python3.10/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.\n",
            "  warnings.warn(msg)\n",
            "Downloading: \"https://download.pytorch.org/models/resnet18-f37072fd.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth\n",
            "100%|██████████| 44.7M/44.7M [00:00<00:00, 379MB/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def train_model(config=None):\n",
        "    with wandb.init(config=config):\n",
        "        # Update the optimizer with the new learning rate\n",
        "        optimizer = optim.SGD(model.fc.parameters(), lr=wandb.config.lr, momentum=0.9)\n",
        "\n",
        "        # Rest of your training and evaluation code here\n",
        "        num_epochs = 10\n",
        "        start_time = time.time()\n",
        "        for epoch in range(num_epochs):\n",
        "            print(f\"Epoch {epoch+1}/{num_epochs}\")\n",
        "            print('-' * 10)\n",
        "\n",
        "            # Training phase\n",
        "            model.train()\n",
        "            running_loss = 0.0\n",
        "\n",
        "            for inputs, labels in train_loader:\n",
        "                inputs = inputs.to(device)\n",
        "                labels = labels.to(device)\n",
        "\n",
        "                optimizer.zero_grad()\n",
        "\n",
        "                # Forward pass\n",
        "                outputs = model(inputs)\n",
        "                loss = criterion(outputs, labels)\n",
        "\n",
        "                # Backward pass and optimization\n",
        "                loss.backward()\n",
        "                optimizer.step()\n",
        "\n",
        "                running_loss += loss.item()\n",
        "\n",
        "            # Validation phase\n",
        "            model.eval()\n",
        "            val_loss = 0.0\n",
        "            correct_preds = 0\n",
        "\n",
        "            with torch.no_grad():\n",
        "                for inputs, labels in val_loader:\n",
        "                    inputs = inputs.to(device)\n",
        "                    labels = labels.to(device)\n",
        "\n",
        "                    # Forward pass\n",
        "                    outputs = model(inputs)\n",
        "                    loss = criterion(outputs, labels)\n",
        "                    val_loss += loss.item()\n",
        "\n",
        "                    # Calculate the number of correct predictions\n",
        "                    _, preds = torch.max(outputs, 1)\n",
        "                    correct_preds += torch.sum(preds == labels.data)\n",
        "\n",
        "            epoch_loss = running_loss / len(train_loader)\n",
        "            epoch_val_loss = val_loss / len(val_loader)\n",
        "            epoch_acc = correct_preds / len(val_loader.dataset)\n",
        "\n",
        "            print('[Test #{}] Loss: {:.4f} Acc: {:.4f}% Time: {:.4f}s'.\n",
        "            format(epoch, epoch_loss, epoch_acc, time.time() - start_time))\n",
        "            # Log the learning rate, accuracy, and loss to wandb\n",
        "            wandb.log({'learning_rate': wandb.config.lr, 'accuracy': epoch_acc, 'loss': epoch_val_loss})\n",
        "\n",
        "    return epoch_acc\n"
      ],
      "metadata": {
        "id": "xYDOYnop5ddI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "wandb.agent(sweep_id, train_model, count=5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000,
          "referenced_widgets": [
            "f6d8c04db4804fb3a9c8cfe0c7a42fb5",
            "06637f43f8d644a59357c3003893ccca",
            "17c9816ee1c0484a8370cbb12e341479",
            "b3dadb2d66b649dcb087dabaa5852b1a",
            "87f45cb23d5a4108944731a102b06c5e",
            "355b80d8e783431fa641c91538825646",
            "c660a6f72b0b4b15aa65a2a13da078a1",
            "6faed737e52d49f7b86c99d4a769d32d",
            "302e935261f5437fbc40b5ce8c78087c",
            "fe3da5e5149c462a99fc531f1640288c",
            "86faad36536743028244debad6438975",
            "6303f379d9f943218da97417430ad13e",
            "b83860649e354f7eac782ee94f23ae88",
            "9872f69a144c481c8e2ed659b1691b58",
            "f6588ccdc1bf46359e16a9b2bbe302ac",
            "a9d7bc2cf3a44e93a328899988cc7281",
            "94b737a9ced84a2581e0ef718175ec35",
            "859ef28efe9f4c8b80c595d2cef02aab",
            "4865f61492c9441696b2ee89d44731e6",
            "6d8f95fcb3904dcdad6944fc10f3781e",
            "36a052297ca44de68977b3ce98e28f24",
            "f4a6630e6a7d47ea91365f76a24ae5b6",
            "02df1027693d49d0b2e07ca236331ff6",
            "b88d9060967a4ffd97a137067398c690"
          ]
        },
        "id": "GmySFSVv1KC1",
        "outputId": "1156ac45-5db6-48e2-81c5-6e804c09c3d0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: jwgxacad with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlr: 0.001\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mmr-srezaei\u001b[0m. Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "VBox(children=(Label(value='Waiting for wandb.init()...\\r'), FloatProgress(value=0.016668456916666703, max=1.0…"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "f6d8c04db4804fb3a9c8cfe0c7a42fb5"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230522_173630-jwgxacad</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/transfer/runs/jwgxacad' target=\"_blank\">northern-sweep-1</a></strong> to <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/transfer/runs/jwgxacad' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/jwgxacad</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "[Test #0] Loss: 1.3640 Acc: 0.5420% Time: 40.7712s\n",
            "Epoch 2/10\n",
            "----------\n",
            "[Test #1] Loss: 1.0427 Acc: 0.5964% Time: 75.4045s\n",
            "Epoch 3/10\n",
            "----------\n",
            "[Test #2] Loss: 0.9500 Acc: 0.6150% Time: 111.1750s\n",
            "Epoch 4/10\n",
            "----------\n",
            "[Test #3] Loss: 0.9167 Acc: 0.6322% Time: 144.9554s\n",
            "Epoch 5/10\n",
            "----------\n",
            "[Test #4] Loss: 0.8841 Acc: 0.6402% Time: 178.5167s\n",
            "Epoch 6/10\n",
            "----------\n",
            "[Test #5] Loss: 0.8696 Acc: 0.6514% Time: 212.3480s\n",
            "Epoch 7/10\n",
            "----------\n",
            "[Test #6] Loss: 0.8503 Acc: 0.6552% Time: 248.2128s\n",
            "Epoch 8/10\n",
            "----------\n",
            "[Test #7] Loss: 0.8344 Acc: 0.6572% Time: 284.1937s\n",
            "Epoch 9/10\n",
            "----------\n",
            "[Test #8] Loss: 0.8209 Acc: 0.6520% Time: 318.1512s\n",
            "Epoch 10/10\n",
            "----------\n",
            "[Test #9] Loss: 0.8171 Acc: 0.6606% Time: 354.7124s\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▁▄▅▆▇▇██▇█</td></tr><tr><td>learning_rate</td><td>▁▁▁▁▁▁▁▁▁▁</td></tr><tr><td>loss</td><td>█▅▃▃▂▂▂▁▁▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.6606</td></tr><tr><td>learning_rate</td><td>0.001</td></tr><tr><td>loss</td><td>0.88675</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">northern-sweep-1</strong> at: <a href='https://wandb.ai/mr-srezaei/transfer/runs/jwgxacad' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/jwgxacad</a><br/>Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230522_173630-jwgxacad/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Sweep Agent: Waiting for job.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Job received.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: 2gq95nov with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlr: 0.0001\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230522_174302-2gq95nov</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/transfer/runs/2gq95nov' target=\"_blank\">bumbling-sweep-2</a></strong> to <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/transfer/runs/2gq95nov' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/2gq95nov</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "[Test #0] Loss: 0.8132 Acc: 0.6634% Time: 36.2410s\n",
            "Epoch 2/10\n",
            "----------\n",
            "[Test #1] Loss: 0.8195 Acc: 0.6618% Time: 69.8526s\n",
            "Epoch 3/10\n",
            "----------\n",
            "[Test #2] Loss: 0.8184 Acc: 0.6650% Time: 103.8597s\n",
            "Epoch 4/10\n",
            "----------\n",
            "[Test #3] Loss: 0.8086 Acc: 0.6654% Time: 142.2045s\n",
            "Epoch 5/10\n",
            "----------\n",
            "[Test #4] Loss: 0.8040 Acc: 0.6652% Time: 176.1389s\n",
            "Epoch 6/10\n",
            "----------\n",
            "[Test #5] Loss: 0.8187 Acc: 0.6652% Time: 210.3880s\n",
            "Epoch 7/10\n",
            "----------\n",
            "[Test #6] Loss: 0.7999 Acc: 0.6624% Time: 245.3717s\n",
            "Epoch 8/10\n",
            "----------\n",
            "[Test #7] Loss: 0.8130 Acc: 0.6652% Time: 280.1540s\n",
            "Epoch 9/10\n",
            "----------\n",
            "[Test #8] Loss: 0.8018 Acc: 0.6664% Time: 314.4616s\n",
            "Epoch 10/10\n",
            "----------\n",
            "[Test #9] Loss: 0.8070 Acc: 0.6660% Time: 349.0051s\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "VBox(children=(Label(value='0.001 MB of 0.001 MB uploaded (0.000 MB deduped)\\r'), FloatProgress(value=1.0, max…"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "302e935261f5437fbc40b5ce8c78087c"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▃▁▆▆▆▆▂▆█▇</td></tr><tr><td>learning_rate</td><td>▁▁▁▁▁▁▁▁▁▁</td></tr><tr><td>loss</td><td>▆▅▂▃▂▃█▃▁▂</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.666</td></tr><tr><td>learning_rate</td><td>0.0001</td></tr><tr><td>loss</td><td>0.87568</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">bumbling-sweep-2</strong> at: <a href='https://wandb.ai/mr-srezaei/transfer/runs/2gq95nov' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/2gq95nov</a><br/>Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230522_174302-2gq95nov/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: 3ark5nrs with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlr: 0.0001\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230522_174911-3ark5nrs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/transfer/runs/3ark5nrs' target=\"_blank\">ruby-sweep-3</a></strong> to <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/transfer/runs/3ark5nrs' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/3ark5nrs</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "[Test #0] Loss: 0.8052 Acc: 0.6662% Time: 36.3905s\n",
            "Epoch 2/10\n",
            "----------\n",
            "[Test #1] Loss: 0.7974 Acc: 0.6656% Time: 72.1323s\n",
            "Epoch 3/10\n",
            "----------\n",
            "[Test #2] Loss: 0.8051 Acc: 0.6650% Time: 107.2112s\n",
            "Epoch 4/10\n",
            "----------\n",
            "[Test #3] Loss: 0.8118 Acc: 0.6638% Time: 141.5506s\n",
            "Epoch 5/10\n",
            "----------\n",
            "[Test #4] Loss: 0.8039 Acc: 0.6684% Time: 177.1074s\n",
            "Epoch 6/10\n",
            "----------\n",
            "[Test #5] Loss: 0.8016 Acc: 0.6674% Time: 212.2513s\n",
            "Epoch 7/10\n",
            "----------\n",
            "[Test #6] Loss: 0.8017 Acc: 0.6660% Time: 246.3060s\n",
            "Epoch 8/10\n",
            "----------\n",
            "[Test #7] Loss: 0.8075 Acc: 0.6662% Time: 280.3745s\n",
            "Epoch 9/10\n",
            "----------\n",
            "[Test #8] Loss: 0.7983 Acc: 0.6678% Time: 318.5867s\n",
            "Epoch 10/10\n",
            "----------\n",
            "[Test #9] Loss: 0.7897 Acc: 0.6684% Time: 352.7271s\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "VBox(children=(Label(value='0.001 MB of 0.001 MB uploaded (0.000 MB deduped)\\r'), FloatProgress(value=1.0, max…"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "94b737a9ced84a2581e0ef718175ec35"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▅▄▃▁█▆▄▅▇█</td></tr><tr><td>learning_rate</td><td>▁▁▁▁▁▁▁▁▁▁</td></tr><tr><td>loss</td><td>▄▅▅█▁▁▆█▁▂</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.6684</td></tr><tr><td>learning_rate</td><td>0.0001</td></tr><tr><td>loss</td><td>0.87142</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">ruby-sweep-3</strong> at: <a href='https://wandb.ai/mr-srezaei/transfer/runs/3ark5nrs' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/3ark5nrs</a><br/>Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230522_174911-3ark5nrs/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: 7gleed1k with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlr: 1e-05\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230522_175522-7gleed1k</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/transfer/runs/7gleed1k' target=\"_blank\">golden-sweep-4</a></strong> to <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/transfer/runs/7gleed1k' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/7gleed1k</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "[Test #0] Loss: 0.7994 Acc: 0.6648% Time: 34.4647s\n",
            "Epoch 2/10\n",
            "----------\n",
            "[Test #1] Loss: 0.8096 Acc: 0.6678% Time: 68.1198s\n",
            "Epoch 3/10\n",
            "----------\n",
            "[Test #2] Loss: 0.7983 Acc: 0.6670% Time: 102.5780s\n",
            "Epoch 4/10\n",
            "----------\n",
            "[Test #3] Loss: 0.8072 Acc: 0.6686% Time: 138.5713s\n",
            "Epoch 5/10\n",
            "----------\n",
            "[Test #4] Loss: 0.7968 Acc: 0.6666% Time: 174.2570s\n",
            "Epoch 6/10\n",
            "----------\n",
            "[Test #5] Loss: 0.8036 Acc: 0.6688% Time: 210.0101s\n",
            "Epoch 7/10\n",
            "----------\n",
            "[Test #6] Loss: 0.8032 Acc: 0.6672% Time: 244.6618s\n",
            "Epoch 8/10\n",
            "----------\n",
            "[Test #7] Loss: 0.8000 Acc: 0.6680% Time: 278.1309s\n",
            "Epoch 9/10\n",
            "----------\n",
            "[Test #8] Loss: 0.7974 Acc: 0.6664% Time: 312.0390s\n",
            "Epoch 10/10\n",
            "----------\n",
            "[Test #9] Loss: 0.7999 Acc: 0.6676% Time: 348.6500s\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>▁▆▅█▄█▅▇▄▆</td></tr><tr><td>learning_rate</td><td>▁▁▁▁▁▁▁▁▁▁</td></tr><tr><td>loss</td><td>█▇▃▂▅▄▅▅▃▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.6676</td></tr><tr><td>learning_rate</td><td>1e-05</td></tr><tr><td>loss</td><td>0.86823</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">golden-sweep-4</strong> at: <a href='https://wandb.ai/mr-srezaei/transfer/runs/7gleed1k' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/7gleed1k</a><br/>Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230522_175522-7gleed1k/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Sweep Agent: Waiting for job.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Job received.\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: Agent Starting Run: wm9b88cp with config:\n",
            "\u001b[34m\u001b[1mwandb\u001b[0m: \tlr: 0.0001\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.3"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20230522_180136-wm9b88cp</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/mr-srezaei/transfer/runs/wm9b88cp' target=\"_blank\">different-sweep-5</a></strong> to <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>Sweep page: <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/mr-srezaei/transfer' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View sweep at <a href='https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/sweeps/yoabtqx9</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/mr-srezaei/transfer/runs/wm9b88cp' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/wm9b88cp</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "[Test #0] Loss: 0.7996 Acc: 0.6706% Time: 33.8878s\n",
            "Epoch 2/10\n",
            "----------\n",
            "[Test #1] Loss: 0.7959 Acc: 0.6672% Time: 69.8209s\n",
            "Epoch 3/10\n",
            "----------\n",
            "[Test #2] Loss: 0.8047 Acc: 0.6678% Time: 103.4074s\n",
            "Epoch 4/10\n",
            "----------\n",
            "[Test #3] Loss: 0.7888 Acc: 0.6688% Time: 139.4800s\n",
            "Epoch 5/10\n",
            "----------\n",
            "[Test #4] Loss: 0.7953 Acc: 0.6666% Time: 173.3399s\n",
            "Epoch 6/10\n",
            "----------\n",
            "[Test #5] Loss: 0.8018 Acc: 0.6698% Time: 207.4189s\n",
            "Epoch 7/10\n",
            "----------\n",
            "[Test #6] Loss: 0.8011 Acc: 0.6680% Time: 242.3962s\n",
            "Epoch 8/10\n",
            "----------\n",
            "[Test #7] Loss: 0.7978 Acc: 0.6704% Time: 277.3398s\n",
            "Epoch 9/10\n",
            "----------\n",
            "[Test #8] Loss: 0.7978 Acc: 0.6676% Time: 311.9802s\n",
            "Epoch 10/10\n",
            "----------\n",
            "[Test #9] Loss: 0.7886 Acc: 0.6688% Time: 348.3799s\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>█▂▃▅▁▇▃█▃▅</td></tr><tr><td>learning_rate</td><td>▁▁▁▁▁▁▁▁▁▁</td></tr><tr><td>loss</td><td>▁█▇▃▆▄▅▃▄▅</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>accuracy</td><td>0.6688</td></tr><tr><td>learning_rate</td><td>0.0001</td></tr><tr><td>loss</td><td>0.86873</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">different-sweep-5</strong> at: <a href='https://wandb.ai/mr-srezaei/transfer/runs/wm9b88cp' target=\"_blank\">https://wandb.ai/mr-srezaei/transfer/runs/wm9b88cp</a><br/>Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20230522_180136-wm9b88cp/logs</code>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Fine Tuning and feature extraction"
      ],
      "metadata": {
        "id": "QpKdws56-xSz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**now we want to measure the accuracy of our model in `fine tuning` process.to do this, first we Update the model to allow fine-tuning ,then Adjust the optimizer to include all model parameters and finally Run the training and evaluation loop as before.**"
      ],
      "metadata": {
        "id": "CFfVN5abQDCG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Step 1: Retrieve logged data\n",
        "api = wandb.Api()\n",
        "runs = api.run(\"mr-srezaei/transfer/wm9b88cp\")\n",
        "\n",
        "if not runs:\n",
        "    print(\"No runs found for the specified project and sweep ID.\")\n",
        "    # Handle this case accordingly\n",
        "\n",
        "\n",
        "# Step 3: Find the best run based on validation loss\n",
        "best_val_loss = float('inf')\n",
        "best_run = None\n",
        "\n",
        "run_data = run.history()\n",
        "Learning_rate = run_data.get('learning_rate')\n",
        "print(run_data)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xIXgLnr1rIdR",
        "outputId": "7f196d03-90a0-4288-bc9a-e48729232ff5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "   learning_rate      loss  _step    _runtime  accuracy    _timestamp\n",
            "0         0.0001  0.864110      0   37.230424    0.6706  1.684779e+09\n",
            "1         0.0001  0.871815      1   73.163257    0.6672  1.684779e+09\n",
            "2         0.0001  0.870255      2  106.749423    0.6678  1.684779e+09\n",
            "3         0.0001  0.865855      3  142.822438    0.6688  1.684779e+09\n",
            "4         0.0001  0.870150      4  176.681074    0.6666  1.684779e+09\n",
            "5         0.0001  0.867737      5  210.760478    0.6698  1.684779e+09\n",
            "6         0.0001  0.868457      6  245.738348    0.6680  1.684779e+09\n",
            "7         0.0001  0.866489      7  280.682201    0.6704  1.684779e+09\n",
            "8         0.0001  0.867560      8  315.321597    0.6676  1.684779e+09\n",
            "9         0.0001  0.868729      9  351.721238    0.6688  1.684779e+09\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Set requires_grad=True for all parameters\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = True\n",
        "\n",
        "# Adjust the optimizer to include all model parameters\n",
        "optimizer = optim.SGD(model.parameters(), lr=Learning_rate[0], momentum=0.9)\n",
        "\n",
        "num_epochs = 10\n",
        "\n",
        "for epoch in range(num_epochs):\n",
        "    print(f\"Epoch {epoch+1}/{num_epochs}\")\n",
        "    print('-' * 10)\n",
        "\n",
        "    # Training phase\n",
        "    model.train()\n",
        "    running_loss = 0.0\n",
        "\n",
        "    for inputs, labels in train_loader:\n",
        "        inputs = inputs.to(device)\n",
        "        labels = labels.to(device)\n",
        "\n",
        "        optimizer.zero_grad()\n",
        "\n",
        "        outputs = model(inputs)\n",
        "        loss = criterion(outputs, labels)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "        running_loss += loss.item()\n",
        "\n",
        "    # Validation phase\n",
        "    model.eval()\n",
        "    val_loss = 0.0\n",
        "    correct_preds = 0\n",
        "\n",
        "    with torch.no_grad():\n",
        "        for inputs, labels in val_loader:\n",
        "            inputs = inputs.to(device)\n",
        "            labels = labels.to(device)\n",
        "\n",
        "            outputs = model(inputs)\n",
        "            loss = criterion(outputs, labels)\n",
        "            val_loss += loss.item()\n",
        "\n",
        "            _, preds = torch.max(outputs, 1)\n",
        "            correct_preds += torch.sum(preds == labels.data)\n",
        "\n",
        "    epoch_loss = running_loss / len(train_loader)\n",
        "    epoch_val_loss = val_loss / len(val_loader)\n",
        "    epoch_acc = correct_preds / len(val_loader.dataset)\n",
        "\n",
        "    # Log the learning rate, accuracy, and loss to wandb\n",
        "    # wandb.log({'learning_rate': Learning_rate[0], 'accuracy': epoch_acc, 'loss': epoch_val_loss})\n",
        "\n",
        "    print(f\"Train Loss: {epoch_loss:.4f}\")\n",
        "    print(f\"Val Loss: {epoch_val_loss:.4f}\")\n",
        "    print(f\"Val Accuracy: {epoch_acc:.4f}\")\n",
        "    print()"
      ],
      "metadata": {
        "id": "x9QpIz5GQZht",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ad48e132-cba6-4a36-bc99-1d19bd8e7a8a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "Train Loss: 0.5966\n",
            "Val Loss: 0.6698\n",
            "Val Accuracy: 0.7610\n",
            "\n",
            "Epoch 2/10\n",
            "----------\n",
            "Train Loss: 0.5871\n",
            "Val Loss: 0.6689\n",
            "Val Accuracy: 0.7592\n",
            "\n",
            "Epoch 3/10\n",
            "----------\n",
            "Train Loss: 0.5789\n",
            "Val Loss: 0.6560\n",
            "Val Accuracy: 0.7656\n",
            "\n",
            "Epoch 4/10\n",
            "----------\n",
            "Train Loss: 0.5673\n",
            "Val Loss: 0.6533\n",
            "Val Accuracy: 0.7648\n",
            "\n",
            "Epoch 5/10\n",
            "----------\n",
            "Train Loss: 0.5692\n",
            "Val Loss: 0.6507\n",
            "Val Accuracy: 0.7690\n",
            "\n",
            "Epoch 6/10\n",
            "----------\n",
            "Train Loss: 0.5603\n",
            "Val Loss: 0.6412\n",
            "Val Accuracy: 0.7724\n",
            "\n",
            "Epoch 7/10\n",
            "----------\n",
            "Train Loss: 0.5547\n",
            "Val Loss: 0.6295\n",
            "Val Accuracy: 0.7760\n",
            "\n",
            "Epoch 8/10\n",
            "----------\n",
            "Train Loss: 0.5465\n",
            "Val Loss: 0.6208\n",
            "Val Accuracy: 0.7788\n",
            "\n",
            "Epoch 9/10\n",
            "----------\n",
            "Train Loss: 0.5266\n",
            "Val Loss: 0.6205\n",
            "Val Accuracy: 0.7786\n",
            "\n",
            "Epoch 10/10\n",
            "----------\n",
            "Train Loss: 0.5345\n",
            "Val Loss: 0.6120\n",
            "Val Accuracy: 0.7850\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# testing phase\n",
        "model.eval()\n",
        "with torch.no_grad():\n",
        "    running_loss = 0.\n",
        "    running_corrects = 0\n",
        "    for inputs, labels in test_loader:\n",
        "        inputs = inputs.to(device)\n",
        "        labels = labels.to(device)\n",
        "        outputs = model(inputs)\n",
        "        _, preds = torch.max(outputs, 1)\n",
        "        loss = criterion(outputs, labels)\n",
        "        running_loss += loss.item() * inputs.size(0)\n",
        "        running_corrects += torch.sum(preds == labels.data)\n",
        "    epoch_loss = running_loss / len(test_data)\n",
        "    epoch_acc = running_corrects / len(test_data) * 100.\n",
        "    print('[Test #{}] Loss: {:.4f} Acc: {:.4f}% '.format(epoch, epoch_loss, epoch_acc))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XSpnkCyV1hdU",
        "outputId": "11d81605-cd0c-4bde-ea79-6fa94b360e59"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Test #5] Loss: 0.6563 Acc: 75.9600% \n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**now we want to measure the model accuracy in `feature extraction` phase.to do this, we first Keep the model parameters frozen for `feature extraction`,then,Adjust the optimizer to only include the parameters of the classifier and finally Run the training and evaluation loop as before.**"
      ],
      "metadata": {
        "id": "iqZIlL7NRLtC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Set requires_grad=False for all parameters except the classifier\n",
        "for name, param in model.named_parameters():\n",
        "    if \"fc\" not in name:\n",
        "        param.requires_grad = False\n",
        "\n",
        "# Adjust the optimizer to only include the parameters of the classifier\n",
        "optimizer = optim.SGD(model.fc.parameters(), lr=Learning_rate[0], momentum=0.9)\n",
        "\n",
        "num_epochs = 10\n",
        "\n",
        "for epoch in range(num_epochs):\n",
        "    print(f\"Epoch {epoch+1}/{num_epochs}\")\n",
        "    print('-' * 10)\n",
        "\n",
        "    # Training phase\n",
        "    model.train()\n",
        "    running_loss = 0.0\n",
        "\n",
        "    for inputs, labels in train_loader:\n",
        "        inputs = inputs.to(device)\n",
        "        labels = labels.to(device)\n",
        "\n",
        "        optimizer.zero_grad()\n",
        "\n",
        "        outputs = model(inputs)\n",
        "        loss = criterion(outputs, labels)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "        running_loss += loss.item()\n",
        "\n",
        "    # Validation phase\n",
        "    model.eval()\n",
        "    val_loss = 0.0\n",
        "    correct_preds = 0\n",
        "\n",
        "    with torch.no_grad():\n",
        "        for inputs, labels in val_loader:\n",
        "            inputs = inputs.to(device)\n",
        "            labels = labels.to(device)\n",
        "\n",
        "            outputs = model(inputs)\n",
        "            loss = criterion(outputs, labels)\n",
        "            val_loss += loss.item()\n",
        "\n",
        "            _, preds = torch.max(outputs, 1)\n",
        "            correct_preds += torch.sum(preds == labels.data)\n",
        "\n",
        "    epoch_loss = running_loss / len(train_loader)\n",
        "    epoch_val_loss = val_loss / len(val_loader)\n",
        "    epoch_acc = correct_preds / len(val_loader.dataset)\n",
        "\n",
        "    # Log the learning rate, accuracy, and loss to wandb\n",
        "    # wandb.log({'learning_rate': learning_rate, 'accuracy': epoch_acc, 'loss': epoch_val_loss})\n",
        "\n",
        "    print(f\"Train Loss: {epoch_loss:.4f}\")\n",
        "    print(f\"Val Loss: {epoch_val_loss:.4f}\")\n",
        "    print(f\"Val Accuracy: {epoch_acc:.4f}\")\n",
        "    print()"
      ],
      "metadata": {
        "id": "P5pFYXYoRL_k",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f258f779-2db8-4115-cd9a-40f80c1abe7f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "----------\n",
            "Train Loss: 0.5240\n",
            "Val Loss: 0.6257\n",
            "Val Accuracy: 0.7806\n",
            "\n",
            "Epoch 2/10\n",
            "----------\n",
            "Train Loss: 0.5258\n",
            "Val Loss: 0.6198\n",
            "Val Accuracy: 0.7816\n",
            "\n",
            "Epoch 3/10\n",
            "----------\n",
            "Train Loss: 0.5292\n",
            "Val Loss: 0.6294\n",
            "Val Accuracy: 0.7798\n",
            "\n",
            "Epoch 4/10\n",
            "----------\n",
            "Train Loss: 0.5247\n",
            "Val Loss: 0.6210\n",
            "Val Accuracy: 0.7790\n",
            "\n",
            "Epoch 5/10\n",
            "----------\n",
            "Train Loss: 0.5302\n",
            "Val Loss: 0.6118\n",
            "Val Accuracy: 0.7860\n",
            "\n",
            "Epoch 6/10\n",
            "----------\n",
            "Train Loss: 0.5315\n",
            "Val Loss: 0.6236\n",
            "Val Accuracy: 0.7826\n",
            "\n",
            "Epoch 7/10\n",
            "----------\n",
            "Train Loss: 0.5284\n",
            "Val Loss: 0.6207\n",
            "Val Accuracy: 0.7824\n",
            "\n",
            "Epoch 8/10\n",
            "----------\n",
            "Train Loss: 0.5380\n",
            "Val Loss: 0.6253\n",
            "Val Accuracy: 0.7804\n",
            "\n",
            "Epoch 9/10\n",
            "----------\n",
            "Train Loss: 0.5356\n",
            "Val Loss: 0.6208\n",
            "Val Accuracy: 0.7794\n",
            "\n",
            "Epoch 10/10\n",
            "----------\n",
            "Train Loss: 0.5331\n",
            "Val Loss: 0.6150\n",
            "Val Accuracy: 0.7820\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# testing phase\n",
        "model.eval()\n",
        "with torch.no_grad():\n",
        "    running_loss = 0.\n",
        "    running_corrects = 0\n",
        "    for inputs, labels in test_loader:\n",
        "        inputs = inputs.to(device)\n",
        "        labels = labels.to(device)\n",
        "        outputs = model(inputs)\n",
        "        _, preds = torch.max(outputs, 1)\n",
        "        loss = criterion(outputs, labels)\n",
        "        running_loss += loss.item() * inputs.size(0)\n",
        "        running_corrects += torch.sum(preds == labels.data)\n",
        "    epoch_loss = running_loss / len(test_data)\n",
        "    epoch_acc = running_corrects / len(test_data) * 100.\n",
        "    print('[Test #{}] Loss: {:.4f} Acc: {:.4f}% '.format(epoch, epoch_loss, epoch_acc))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jdtrf2Xh2hsz",
        "outputId": "23c30eb4-14c5-4415-e9b0-a2d2063dfdbc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Test #9] Loss: 0.6583 Acc: 75.7200% \n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Evaluating Files\n",
        "in the code snippets below we save our evaluating scores into csv file"
      ],
      "metadata": {
        "id": "ChYPm8qx75Dh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install wandb"
      ],
      "metadata": {
        "id": "exg4OufQ75Dq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import wandb\n",
        "import tensorflow as tf\n",
        "wandb.login()\n",
        "\n",
        "run = wandb.init()\n",
        "artifact = run.use_artifact('mr-srezaei/transfer/run-wm9b88cp-history:v0', type='wandb-history')\n",
        "artifact_dir = artifact.download()\n",
        "\n",
        "best_model = tf.keras.models.load_model(artifact_dir)"
      ],
      "metadata": {
        "id": "4OknJiC_75Dr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from os import listdir\n",
        "from os.path import isfile, join, splitext\n",
        "from tqdm import tqdm\n",
        "import numpy as np\n",
        "import csv\n",
        "import pandas as pd\n",
        "\n",
        "# Define your classes\n",
        "classes = ['articulated_truck', 'bus', 'car', 'work_van', 'background']\n",
        "\n",
        "def load_and_preprocess_image(image_path):\n",
        "    # Read image from file\n",
        "    image = tf.io.read_file(image_path)\n",
        "    \n",
        "    # Decode image and convert to desired format\n",
        "    image = tf.image.decode_image(image, channels=3)\n",
        "    \n",
        "    # Resize the image to (128, 128)\n",
        "    image = tf.image.resize(image, [128, 128])\n",
        "    \n",
        "    # Rescale the pixel values to the range [0, 1]\n",
        "    image = image / 255.0\n",
        "    \n",
        "    return image\n",
        "\n",
        "def classify_image(path_to_image):\n",
        "\n",
        "    # Load the image and preprocess it if needed\n",
        "    image = load_and_preprocess_image(path_to_image)\n",
        "\n",
        "    # Make predictions using your trained model\n",
        "    predictions = best_model.predict(np.array([image]))  # Assuming 'model' is your trained model\n",
        "\n",
        "    # Get the predicted label\n",
        "    label_index = np.argmax(predictions[0])\n",
        "    label = classes[label_index]\n",
        "\n",
        "    return label\n",
        "\n",
        "def parse_dataset(path_to_dataset):\n",
        "    '''\n",
        "    Parse every image contained in 'path_to_dataset' (a path to the training\n",
        "    or testing set), classify each image and save in a csv file the resulting\n",
        "    assignment\n",
        "\n",
        "    dataset_result: dict structure returned by the function.  It contains the\n",
        "            label of each image\n",
        "    '''\n",
        "    llist = listdir(path_to_dataset)\n",
        "    dataset_result = {}\n",
        "\n",
        "    for name in tqdm(llist):\n",
        "        dn = join(path_to_dataset, name)\n",
        "        if isfile(dn):\n",
        "            label = classify_image(dn)\n",
        "            file_nb, file_ext = splitext(name)\n",
        "            dataset_result[file_nb] = label\n",
        "\n",
        "        else:\n",
        "            file_list = listdir(dn)\n",
        "            for file_name in file_list:\n",
        "                file_name_with_path = join(dn, file_name)\n",
        "                label = classify_image(file_name_with_path)\n",
        "                file_nb, file_ext = splitext(file_name)\n",
        "                if file_nb in dataset_result.keys():\n",
        "                    print('error! ', file_nb, dataset_result[file_nb], ' vs ', file_name_with_path)\n",
        "                dataset_result[file_nb] = label\n",
        "\n",
        "    return dataset_result\n",
        "\n",
        "def save_classification_result(dataset_result, output_csv_file_name):\n",
        "    '''\n",
        "    save the dataset_result (a dict structure containing the class of every image)\n",
        "    into a valid csv file.\n",
        "    '''\n",
        "\n",
        "    with open(output_csv_file_name, 'w', newline='') as csvfile:\n",
        "        fieldnames = ['file_name', 'class_label']\n",
        "        writer = csv.DictWriter(csvfile, fieldnames=fieldnames)\n",
        "        writer.writeheader()\n",
        "\n",
        "        for file_name, class_label in dataset_result.items():\n",
        "            file_name_without_zeros = file_name.lstrip('0')\n",
        "            writer.writerow({'file_name': file_name_without_zeros, 'class_label': class_label})\n",
        "\n",
        "    csvfile.close()\n",
        "\n",
        "if __name__ == '__main__':\n",
        "    # Specify the path to the test dataset and the output CSV file name\n",
        "    path_to_test_dataset = '/content/sample_data/miotcd/HW02-Practical/Dataset/test'  # Replace with your actual path\n",
        "    output_csv_file_name = 'your_results_transfer_test.csv'  # Replace with your desired output file name\n",
        "\n",
        "    dataset_result = parse_dataset(path_to_test_dataset)\n",
        "    save_classification_result(dataset_result, output_csv_file_name)"
      ],
      "metadata": {
        "id": "AI6-A3_-75Dr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import csv\n",
        "from sklearn.metrics import confusion_matrix, cohen_kappa_score, precision_score, recall_score, f1_score\n",
        "import numpy as np\n",
        "\n",
        "def compute_metric(result, gt, classes):\n",
        "    y_true = []\n",
        "    y_pred = []\n",
        "\n",
        "\n",
        "    for img in gt.keys():\n",
        "        if img not in result:\n",
        "            raise ValueError(f\"Image '{img}' missing in the result data.\")\n",
        "        \n",
        "        y_true.append(gt[img])\n",
        "        y_pred.append(result[img])\n",
        "\n",
        "    cm = confusion_matrix(y_true, y_pred, labels=classes)\n",
        "    cks = cohen_kappa_score(y_true, y_pred, labels=classes)\n",
        "    ps = precision_score(y_true, y_pred, average=None)\n",
        "    rs = recall_score(y_true, y_pred, average=None)\n",
        "    fs = f1_score(y_true, y_pred, average=None)\n",
        "\n",
        "    metrics = {}\n",
        "    metrics['confusion matrix'] = cm\n",
        "    metrics['cohen kappa score'] = cks\n",
        "    metrics['precision score'] = ps\n",
        "    metrics['recall score'] = rs\n",
        "    metrics['f1 score'] = fs\n",
        "    metrics['mean recall'] = np.mean(rs)\n",
        "    metrics['mean precision'] = np.mean(ps)\n",
        "    metrics['accuracy'] = np.diagonal(cm).sum() / float(len(y_true))\n",
        "    return metrics\n",
        "\n",
        "def csv_evaluation(gt_file, res_file, classes):\n",
        "    gt = {}\n",
        "    with open(gt_file, 'r') as csv_file:\n",
        "        reader = csv.reader(csv_file, delimiter=',')\n",
        "        for row in reader:\n",
        "            gt[row[0]] = row[1]\n",
        "\n",
        "\n",
        "    results = {}\n",
        "    with open(res_file, 'r') as csv_file:\n",
        "        reader = csv.reader(csv_file, delimiter=',')\n",
        "        for row in reader:\n",
        "            results[row[0]] = row[1]\n",
        "\n",
        "\n",
        "    metrics = compute_metric(results, gt, classes)\n",
        "    return metrics\n",
        "\n",
        "if __name__ == '__main__':\n",
        "    gt_file = '/content/drive/MyDrive/Colab Notebooks/HW-2/gt_test.csv'\n",
        "    res_file = '/content/drive/MyDrive/Colab Notebooks/HW-2/your_results_test.csv'\n",
        "    classes = ['articulated_truck', 'bus', 'car', 'work_van', 'background']  # Provide the list of classes\n",
        "\n",
        "    print('Computing score between', gt_file, 'and', res_file, '\\n')\n",
        "    metrics = csv_evaluation(gt_file, res_file, classes)\n",
        "\n",
        "    for key in metrics.keys():\n",
        "        print(key + ':')\n",
        "        print(metrics[key])\n",
        "        print('\\n')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bb2ca09f-3a97-428e-a2a6-56eb92666663",
        "id": "H0mQOAjo75Ds"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Computing score between /content/drive/MyDrive/Colab Notebooks/HW-2/gt_test.csv and /content/drive/MyDrive/Colab Notebooks/HW-2/your_results_test.csv \n",
            "\n",
            "confusion matrix:\n",
            "[[875  46  28  18  33]\n",
            " [ 40   9 913   9  29]\n",
            " [ 53  20  33 738 156]\n",
            " [ 51  61  36  63 789]\n",
            " [ 10 977   2   7   4]]\n",
            "\n",
            "\n",
            "cohen kappa score:\n",
            "-0.0040000000000000036\n",
            "\n",
            "\n",
            "precision score:\n",
            "[0.85034014 0.00395648 0.00808625 0.0326087  0.0754491 ]\n",
            "\n",
            "\n",
            "recall score:\n",
            "[0.875 0.004 0.009 0.033 0.063]\n",
            "\n",
            "\n",
            "f1 score:\n",
            "[0.86249384 0.00397812 0.00851869 0.03280318 0.06866485]\n",
            "\n",
            "\n",
            "mean recall:\n",
            "0.1968\n",
            "\n",
            "\n",
            "mean precision:\n",
            "0.19408813312124037\n",
            "\n",
            "\n",
            "accuracy:\n",
            "0.1968\n",
            "\n",
            "\n"
          ]
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "collapsed_sections": [
        "tE7CuOwvpSlr",
        "bei6FmA1-80Q",
        "vZB64EuwEeaH",
        "iL-Gp126K7_6",
        "A2TkENIEpvmF",
        "A4WLFMQb-nO1",
        "QpKdws56-xSz",
        "ChYPm8qx75Dh"
      ],
      "include_colab_link": true
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "16e7eaf142c24de298ef72bb74da9521": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "VBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "VBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "VBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_ca6439a995574a83bd4131583127f6e8",
              "IPY_MODEL_d8a64efba9694880b23931ea7ac3dd7b"
            ],
            "layout": "IPY_MODEL_3af53fc5aa4a470abe230aef46cdf584"
          }
        },
        "ca6439a995574a83bd4131583127f6e8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "LabelModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "LabelModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "LabelView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_1adcce1812a84dffa6e510f1d556644d",
            "placeholder": "​",
            "style": "IPY_MODEL_ce71f5b3dd164f6683cc62cc6931688b",
            "value": "3300.167 MB of 3300.167 MB uploaded (0.205 MB deduped)\r"
          }
        },
        "d8a64efba9694880b23931ea7ac3dd7b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_69604f7275b34bdca459ccf439217697",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_ff576fc9199443f491b3b9f3e20ec943",
            "value": 1
          }
        },
        "3af53fc5aa4a470abe230aef46cdf584": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "1adcce1812a84dffa6e510f1d556644d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ce71f5b3dd164f6683cc62cc6931688b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "69604f7275b34bdca459ccf439217697": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ff576fc9199443f491b3b9f3e20ec943": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "f6d8c04db4804fb3a9c8cfe0c7a42fb5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "VBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "VBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "VBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_06637f43f8d644a59357c3003893ccca",
              "IPY_MODEL_17c9816ee1c0484a8370cbb12e341479"
            ],
            "layout": "IPY_MODEL_b3dadb2d66b649dcb087dabaa5852b1a"
          }
        },
        "06637f43f8d644a59357c3003893ccca": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "LabelModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "LabelModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "LabelView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_87f45cb23d5a4108944731a102b06c5e",
            "placeholder": "​",
            "style": "IPY_MODEL_355b80d8e783431fa641c91538825646",
            "value": "Waiting for wandb.init()...\r"
          }
        },
        "17c9816ee1c0484a8370cbb12e341479": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c660a6f72b0b4b15aa65a2a13da078a1",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_6faed737e52d49f7b86c99d4a769d32d",
            "value": 1
          }
        },
        "b3dadb2d66b649dcb087dabaa5852b1a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "87f45cb23d5a4108944731a102b06c5e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "355b80d8e783431fa641c91538825646": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "c660a6f72b0b4b15aa65a2a13da078a1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "6faed737e52d49f7b86c99d4a769d32d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "302e935261f5437fbc40b5ce8c78087c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "VBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "VBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "VBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_fe3da5e5149c462a99fc531f1640288c",
              "IPY_MODEL_86faad36536743028244debad6438975"
            ],
            "layout": "IPY_MODEL_6303f379d9f943218da97417430ad13e"
          }
        },
        "fe3da5e5149c462a99fc531f1640288c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "LabelModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "LabelModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "LabelView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b83860649e354f7eac782ee94f23ae88",
            "placeholder": "​",
            "style": "IPY_MODEL_9872f69a144c481c8e2ed659b1691b58",
            "value": "0.010 MB of 0.010 MB uploaded (0.000 MB deduped)\r"
          }
        },
        "86faad36536743028244debad6438975": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_f6588ccdc1bf46359e16a9b2bbe302ac",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_a9d7bc2cf3a44e93a328899988cc7281",
            "value": 1
          }
        },
        "6303f379d9f943218da97417430ad13e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b83860649e354f7eac782ee94f23ae88": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "9872f69a144c481c8e2ed659b1691b58": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "f6588ccdc1bf46359e16a9b2bbe302ac": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a9d7bc2cf3a44e93a328899988cc7281": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "94b737a9ced84a2581e0ef718175ec35": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "VBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "VBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "VBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_859ef28efe9f4c8b80c595d2cef02aab",
              "IPY_MODEL_4865f61492c9441696b2ee89d44731e6"
            ],
            "layout": "IPY_MODEL_6d8f95fcb3904dcdad6944fc10f3781e"
          }
        },
        "859ef28efe9f4c8b80c595d2cef02aab": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "LabelModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "LabelModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "LabelView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_36a052297ca44de68977b3ce98e28f24",
            "placeholder": "​",
            "style": "IPY_MODEL_f4a6630e6a7d47ea91365f76a24ae5b6",
            "value": "0.010 MB of 0.010 MB uploaded (0.000 MB deduped)\r"
          }
        },
        "4865f61492c9441696b2ee89d44731e6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_02df1027693d49d0b2e07ca236331ff6",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_b88d9060967a4ffd97a137067398c690",
            "value": 1
          }
        },
        "6d8f95fcb3904dcdad6944fc10f3781e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "36a052297ca44de68977b3ce98e28f24": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "f4a6630e6a7d47ea91365f76a24ae5b6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "02df1027693d49d0b2e07ca236331ff6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b88d9060967a4ffd97a137067398c690": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}