Method_2_Kaggle_Notebook.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Method 2: Kaggle Notebook.ipynb",
      "provenance": [],
      "machine_shape": "hm",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/thanhtrinh2003/GarbageClassificationKaggle/blob/main/Method_2_Kaggle_Notebook.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "64HsWLS_REd2",
        "outputId": "53ba4dbe-e2f7-42fe-958f-56308203e3d6"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NFIX5ax9aadj",
        "outputId": "07591a78-08f6-43ff-cb0c-7c9e0350a549"
      },
      "source": [
        "import re\n",
        "import os\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.applications import ResNet50\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "\n",
        "def list_dataset():\n",
        "    for dirname, _, filenames in os.walk('/content/drive/MyDrive/Lumiere 2021/Code/Garbage classification/'):\n",
        "        for filename in filenames:\n",
        "            print(os.path.join(dirname, filename))\n",
        "\n",
        "            \n",
        "# Add class name prefix to each path based on class name include in filename\n",
        "def add_class_name_prefix(df, col_name):\n",
        "    df[col_name] = df[col_name].apply(lambda x: x[:re.search(\"\\d\",x).start()] + '/' + x)\n",
        "    return df\n",
        "\n",
        "\n",
        "def class_id_to_label(id):\n",
        "    label_map = {1: 'glass', 2: 'paper', 3: 'cardboard', 4: 'plastic', 5: 'metal', 6: 'trash'}\n",
        "    return label_map[id]\n",
        "\n",
        "IMAGES_DIR = '/content/drive/MyDrive/Lumiere 2021/Code/Garbage classification/'\n",
        "    \n",
        "train_file = '/content/drive/MyDrive/Lumiere 2021/Code/Info/one-indexed-files-notrash_train.txt'\n",
        "val_file   = '/content/drive/MyDrive/Lumiere 2021/Code/Info/one-indexed-files-notrash_val.txt'\n",
        "test_file  = '/content/drive/MyDrive/Lumiere 2021/Code/Info/one-indexed-files-notrash_test.txt'\n",
        "\n",
        "df_train = pd.read_csv(train_file, sep=' ', header=None, names=['rel_path', 'label'])\n",
        "df_valid = pd.read_csv(val_file,   sep=' ', header=None, names=['rel_path', 'label'])\n",
        "df_test  = pd.read_csv(val_file,   sep=' ', header=None, names=['rel_path', 'label'])\n",
        "\n",
        "df_train = add_class_name_prefix(df_train, 'rel_path')\n",
        "df_valid = add_class_name_prefix(df_valid, 'rel_path')\n",
        "df_test  = add_class_name_prefix(df_test,  'rel_path')\n",
        "\n",
        "df_train['label'] = df_train['label'].apply(class_id_to_label)\n",
        "df_valid['label'] = df_valid['label'].apply(class_id_to_label)\n",
        "df_test['label']  = df_test['label'].apply(class_id_to_label)\n",
        "\n",
        "print(f'Found {len(df_train)} training, {len(df_valid)} validation and {len(df_test)} samples.')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 1768 training, 328 validation and 328 samples.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 202
        },
        "id": "X2RccNUPbo7o",
        "outputId": "22aa3101-4263-478c-82a2-a1aa284643c1"
      },
      "source": [
        "df_train.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>rel_path</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>cardboard/cardboard202.jpg</td>\n",
              "      <td>cardboard</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>paper/paper472.jpg</td>\n",
              "      <td>paper</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>paper/paper522.jpg</td>\n",
              "      <td>paper</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>glass/glass189.jpg</td>\n",
              "      <td>glass</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>glass/glass325.jpg</td>\n",
              "      <td>glass</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                     rel_path      label\n",
              "0  cardboard/cardboard202.jpg  cardboard\n",
              "1          paper/paper472.jpg      paper\n",
              "2          paper/paper522.jpg      paper\n",
              "3          glass/glass189.jpg      glass\n",
              "4          glass/glass325.jpg      glass"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BwkgZ4NDyjZe",
        "outputId": "2f36a86e-6096-4e3b-cd82-5903066e1683"
      },
      "source": [
        "datagen = ImageDataGenerator()\n",
        "\n",
        "datagen_train = datagen.flow_from_dataframe(\n",
        "    dataframe=df_train,\n",
        "    directory=IMAGES_DIR,\n",
        "    x_col='rel_path',\n",
        "    y_col='label',\n",
        "    color_mode=\"rgb\",\n",
        "    class_mode=\"categorical\",\n",
        "    batch_size=32,\n",
        "    shuffle=True,\n",
        "    seed=7,\n",
        ")\n",
        "\n",
        "datagen_valid = datagen.flow_from_dataframe(\n",
        "    dataframe=df_valid,\n",
        "    directory=IMAGES_DIR,\n",
        "    x_col='rel_path',\n",
        "    y_col='label',\n",
        "    color_mode=\"rgb\",\n",
        "    class_mode=\"categorical\",\n",
        "    batch_size=32,\n",
        "    shuffle=True,\n",
        "    seed=7,\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 1768 validated image filenames belonging to 6 classes.\n",
            "Found 328 validated image filenames belonging to 6 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sj6rdaGRynnl",
        "outputId": "074157a9-9285-4aed-a830-5b3a5cc61f0a"
      },
      "source": [
        "def build_model(num_classes):\n",
        "    base_model = ResNet50(weights='imagenet', include_top=False)\n",
        "\n",
        "    x = base_model.output\n",
        "    x = tf.keras.layers.GlobalAveragePooling2D()(x)\n",
        "    x = tf.keras.layers.Dense(1024, activation='relu')(x)\n",
        "    predictions = tf.keras.layers.Dense(num_classes, activation='softmax')(x)\n",
        "\n",
        "    model = tf.keras.Model(inputs=base_model.input, outputs=predictions)\n",
        "\n",
        "    for layer in base_model.layers:\n",
        "        layer.trainable = False\n",
        "        \n",
        "    return model\n",
        "\n",
        "\n",
        "net = build_model(num_classes=6)\n",
        "\n",
        "net.compile(optimizer='Adam',\n",
        "            loss='categorical_crossentropy',\n",
        "            metrics=[tf.keras.metrics.categorical_accuracy])\n",
        "\n",
        "net.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "94773248/94765736 [==============================] - 1s 0us/step\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            [(None, None, None,  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "conv1_pad (ZeroPadding2D)       (None, None, None, 3 0           input_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "conv1_conv (Conv2D)             (None, None, None, 6 9472        conv1_pad[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv1_bn (BatchNormalization)   (None, None, None, 6 256         conv1_conv[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1_relu (Activation)         (None, None, None, 6 0           conv1_bn[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "pool1_pad (ZeroPadding2D)       (None, None, None, 6 0           conv1_relu[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "pool1_pool (MaxPooling2D)       (None, None, None, 6 0           pool1_pad[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_1_conv (Conv2D)    (None, None, None, 6 4160        pool1_pool[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_1_bn (BatchNormali (None, None, None, 6 256         conv2_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_1_relu (Activation (None, None, None, 6 0           conv2_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_2_conv (Conv2D)    (None, None, None, 6 36928       conv2_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_2_bn (BatchNormali (None, None, None, 6 256         conv2_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_2_relu (Activation (None, None, None, 6 0           conv2_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_0_conv (Conv2D)    (None, None, None, 2 16640       pool1_pool[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_3_conv (Conv2D)    (None, None, None, 2 16640       conv2_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_0_bn (BatchNormali (None, None, None, 2 1024        conv2_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_3_bn (BatchNormali (None, None, None, 2 1024        conv2_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_add (Add)          (None, None, None, 2 0           conv2_block1_0_bn[0][0]          \n",
            "                                                                 conv2_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block1_out (Activation)   (None, None, None, 2 0           conv2_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_1_conv (Conv2D)    (None, None, None, 6 16448       conv2_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_1_bn (BatchNormali (None, None, None, 6 256         conv2_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_1_relu (Activation (None, None, None, 6 0           conv2_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_2_conv (Conv2D)    (None, None, None, 6 36928       conv2_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_2_bn (BatchNormali (None, None, None, 6 256         conv2_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_2_relu (Activation (None, None, None, 6 0           conv2_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_3_conv (Conv2D)    (None, None, None, 2 16640       conv2_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_3_bn (BatchNormali (None, None, None, 2 1024        conv2_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_add (Add)          (None, None, None, 2 0           conv2_block1_out[0][0]           \n",
            "                                                                 conv2_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block2_out (Activation)   (None, None, None, 2 0           conv2_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_1_conv (Conv2D)    (None, None, None, 6 16448       conv2_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_1_bn (BatchNormali (None, None, None, 6 256         conv2_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_1_relu (Activation (None, None, None, 6 0           conv2_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_2_conv (Conv2D)    (None, None, None, 6 36928       conv2_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_2_bn (BatchNormali (None, None, None, 6 256         conv2_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_2_relu (Activation (None, None, None, 6 0           conv2_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_3_conv (Conv2D)    (None, None, None, 2 16640       conv2_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_3_bn (BatchNormali (None, None, None, 2 1024        conv2_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_add (Add)          (None, None, None, 2 0           conv2_block2_out[0][0]           \n",
            "                                                                 conv2_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv2_block3_out (Activation)   (None, None, None, 2 0           conv2_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_1_conv (Conv2D)    (None, None, None, 1 32896       conv2_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_1_bn (BatchNormali (None, None, None, 1 512         conv3_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_1_relu (Activation (None, None, None, 1 0           conv3_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_2_bn (BatchNormali (None, None, None, 1 512         conv3_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_2_relu (Activation (None, None, None, 1 0           conv3_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_0_conv (Conv2D)    (None, None, None, 5 131584      conv2_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_0_bn (BatchNormali (None, None, None, 5 2048        conv3_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_add (Add)          (None, None, None, 5 0           conv3_block1_0_bn[0][0]          \n",
            "                                                                 conv3_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block1_out (Activation)   (None, None, None, 5 0           conv3_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_1_conv (Conv2D)    (None, None, None, 1 65664       conv3_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_1_bn (BatchNormali (None, None, None, 1 512         conv3_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_1_relu (Activation (None, None, None, 1 0           conv3_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_2_bn (BatchNormali (None, None, None, 1 512         conv3_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_2_relu (Activation (None, None, None, 1 0           conv3_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_add (Add)          (None, None, None, 5 0           conv3_block1_out[0][0]           \n",
            "                                                                 conv3_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block2_out (Activation)   (None, None, None, 5 0           conv3_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_1_conv (Conv2D)    (None, None, None, 1 65664       conv3_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_1_bn (BatchNormali (None, None, None, 1 512         conv3_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_1_relu (Activation (None, None, None, 1 0           conv3_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_2_bn (BatchNormali (None, None, None, 1 512         conv3_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_2_relu (Activation (None, None, None, 1 0           conv3_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_add (Add)          (None, None, None, 5 0           conv3_block2_out[0][0]           \n",
            "                                                                 conv3_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block3_out (Activation)   (None, None, None, 5 0           conv3_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_1_conv (Conv2D)    (None, None, None, 1 65664       conv3_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_1_bn (BatchNormali (None, None, None, 1 512         conv3_block4_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_1_relu (Activation (None, None, None, 1 0           conv3_block4_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_2_conv (Conv2D)    (None, None, None, 1 147584      conv3_block4_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_2_bn (BatchNormali (None, None, None, 1 512         conv3_block4_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_2_relu (Activation (None, None, None, 1 0           conv3_block4_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_3_conv (Conv2D)    (None, None, None, 5 66048       conv3_block4_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_3_bn (BatchNormali (None, None, None, 5 2048        conv3_block4_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_add (Add)          (None, None, None, 5 0           conv3_block3_out[0][0]           \n",
            "                                                                 conv3_block4_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv3_block4_out (Activation)   (None, None, None, 5 0           conv3_block4_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_1_conv (Conv2D)    (None, None, None, 2 131328      conv3_block4_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_1_relu (Activation (None, None, None, 2 0           conv4_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_2_relu (Activation (None, None, None, 2 0           conv4_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_0_conv (Conv2D)    (None, None, None, 1 525312      conv3_block4_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_0_bn (BatchNormali (None, None, None, 1 4096        conv4_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_add (Add)          (None, None, None, 1 0           conv4_block1_0_bn[0][0]          \n",
            "                                                                 conv4_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block1_out (Activation)   (None, None, None, 1 0           conv4_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_1_relu (Activation (None, None, None, 2 0           conv4_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_2_relu (Activation (None, None, None, 2 0           conv4_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_add (Add)          (None, None, None, 1 0           conv4_block1_out[0][0]           \n",
            "                                                                 conv4_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block2_out (Activation)   (None, None, None, 1 0           conv4_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_1_relu (Activation (None, None, None, 2 0           conv4_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_2_relu (Activation (None, None, None, 2 0           conv4_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_add (Add)          (None, None, None, 1 0           conv4_block2_out[0][0]           \n",
            "                                                                 conv4_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block3_out (Activation)   (None, None, None, 1 0           conv4_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block4_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_1_relu (Activation (None, None, None, 2 0           conv4_block4_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block4_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block4_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_2_relu (Activation (None, None, None, 2 0           conv4_block4_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block4_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block4_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_add (Add)          (None, None, None, 1 0           conv4_block3_out[0][0]           \n",
            "                                                                 conv4_block4_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block4_out (Activation)   (None, None, None, 1 0           conv4_block4_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block4_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block5_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_1_relu (Activation (None, None, None, 2 0           conv4_block5_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block5_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block5_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_2_relu (Activation (None, None, None, 2 0           conv4_block5_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block5_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block5_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_add (Add)          (None, None, None, 1 0           conv4_block4_out[0][0]           \n",
            "                                                                 conv4_block5_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block5_out (Activation)   (None, None, None, 1 0           conv4_block5_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_1_conv (Conv2D)    (None, None, None, 2 262400      conv4_block5_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_1_bn (BatchNormali (None, None, None, 2 1024        conv4_block6_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_1_relu (Activation (None, None, None, 2 0           conv4_block6_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_2_conv (Conv2D)    (None, None, None, 2 590080      conv4_block6_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_2_bn (BatchNormali (None, None, None, 2 1024        conv4_block6_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_2_relu (Activation (None, None, None, 2 0           conv4_block6_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_3_conv (Conv2D)    (None, None, None, 1 263168      conv4_block6_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_3_bn (BatchNormali (None, None, None, 1 4096        conv4_block6_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_add (Add)          (None, None, None, 1 0           conv4_block5_out[0][0]           \n",
            "                                                                 conv4_block6_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv4_block6_out (Activation)   (None, None, None, 1 0           conv4_block6_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_1_conv (Conv2D)    (None, None, None, 5 524800      conv4_block6_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_1_bn (BatchNormali (None, None, None, 5 2048        conv5_block1_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_1_relu (Activation (None, None, None, 5 0           conv5_block1_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_2_conv (Conv2D)    (None, None, None, 5 2359808     conv5_block1_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_2_bn (BatchNormali (None, None, None, 5 2048        conv5_block1_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_2_relu (Activation (None, None, None, 5 0           conv5_block1_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_0_conv (Conv2D)    (None, None, None, 2 2099200     conv4_block6_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_3_conv (Conv2D)    (None, None, None, 2 1050624     conv5_block1_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_0_bn (BatchNormali (None, None, None, 2 8192        conv5_block1_0_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_3_bn (BatchNormali (None, None, None, 2 8192        conv5_block1_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_add (Add)          (None, None, None, 2 0           conv5_block1_0_bn[0][0]          \n",
            "                                                                 conv5_block1_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block1_out (Activation)   (None, None, None, 2 0           conv5_block1_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_1_conv (Conv2D)    (None, None, None, 5 1049088     conv5_block1_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_1_bn (BatchNormali (None, None, None, 5 2048        conv5_block2_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_1_relu (Activation (None, None, None, 5 0           conv5_block2_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_2_conv (Conv2D)    (None, None, None, 5 2359808     conv5_block2_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_2_bn (BatchNormali (None, None, None, 5 2048        conv5_block2_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_2_relu (Activation (None, None, None, 5 0           conv5_block2_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_3_conv (Conv2D)    (None, None, None, 2 1050624     conv5_block2_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_3_bn (BatchNormali (None, None, None, 2 8192        conv5_block2_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_add (Add)          (None, None, None, 2 0           conv5_block1_out[0][0]           \n",
            "                                                                 conv5_block2_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block2_out (Activation)   (None, None, None, 2 0           conv5_block2_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_1_conv (Conv2D)    (None, None, None, 5 1049088     conv5_block2_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_1_bn (BatchNormali (None, None, None, 5 2048        conv5_block3_1_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_1_relu (Activation (None, None, None, 5 0           conv5_block3_1_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_2_conv (Conv2D)    (None, None, None, 5 2359808     conv5_block3_1_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_2_bn (BatchNormali (None, None, None, 5 2048        conv5_block3_2_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_2_relu (Activation (None, None, None, 5 0           conv5_block3_2_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_3_conv (Conv2D)    (None, None, None, 2 1050624     conv5_block3_2_relu[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_3_bn (BatchNormali (None, None, None, 2 8192        conv5_block3_3_conv[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_add (Add)          (None, None, None, 2 0           conv5_block2_out[0][0]           \n",
            "                                                                 conv5_block3_3_bn[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv5_block3_out (Activation)   (None, None, None, 2 0           conv5_block3_add[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d (Globa (None, 2048)         0           conv5_block3_out[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "dense (Dense)                   (None, 1024)         2098176     global_average_pooling2d[0][0]   \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 6)            6150        dense[0][0]                      \n",
            "==================================================================================================\n",
            "Total params: 25,692,038\n",
            "Trainable params: 2,104,326\n",
            "Non-trainable params: 23,587,712\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ia1dcjNTjwsi",
        "outputId": "a6fd5451-dc03-47a0-8ddb-f2e2e7ddb140"
      },
      "source": [
        "count = 0;\n",
        "for x in net.layers:\n",
        "  if x.trainable == False:\n",
        "    count = count +1\n",
        "\n",
        "print(count)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "175\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KYYIX2X8jwqn"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aV58QWAV0Yql",
        "outputId": "33a129df-10e2-4fa1-e4e4-929953cc68ea"
      },
      "source": [
        "early_stop = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5, verbose=1, restore_best_weights=True)\n",
        "\n",
        "history = net.fit_generator(\n",
        "    generator=datagen_train,\n",
        "    validation_data=datagen_valid,\n",
        "    epochs=30,\n",
        "    validation_freq=1,\n",
        "    callbacks=[early_stop]\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:1940: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
            "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/30\n",
            "56/56 [==============================] - 1231s 22s/step - loss: 1.0404 - categorical_accuracy: 0.6900 - val_loss: 0.4487 - val_categorical_accuracy: 0.8323\n",
            "Epoch 2/30\n",
            "56/56 [==============================] - 10s 171ms/step - loss: 0.3279 - categorical_accuracy: 0.8891 - val_loss: 0.3329 - val_categorical_accuracy: 0.8720\n",
            "Epoch 3/30\n",
            "56/56 [==============================] - 10s 171ms/step - loss: 0.1965 - categorical_accuracy: 0.9367 - val_loss: 0.4029 - val_categorical_accuracy: 0.8598\n",
            "Epoch 4/30\n",
            "56/56 [==============================] - 10s 173ms/step - loss: 0.1635 - categorical_accuracy: 0.9457 - val_loss: 0.4018 - val_categorical_accuracy: 0.8537\n",
            "Epoch 5/30\n",
            "56/56 [==============================] - 10s 169ms/step - loss: 0.0953 - categorical_accuracy: 0.9712 - val_loss: 0.4033 - val_categorical_accuracy: 0.8537\n",
            "Epoch 6/30\n",
            "56/56 [==============================] - 10s 170ms/step - loss: 0.0642 - categorical_accuracy: 0.9768 - val_loss: 0.2641 - val_categorical_accuracy: 0.9024\n",
            "Epoch 7/30\n",
            "56/56 [==============================] - 10s 173ms/step - loss: 0.0235 - categorical_accuracy: 0.9983 - val_loss: 0.2762 - val_categorical_accuracy: 0.9055\n",
            "Epoch 8/30\n",
            "56/56 [==============================] - 10s 173ms/step - loss: 0.0152 - categorical_accuracy: 0.9983 - val_loss: 0.2786 - val_categorical_accuracy: 0.9024\n",
            "Epoch 9/30\n",
            "56/56 [==============================] - 10s 172ms/step - loss: 0.0106 - categorical_accuracy: 0.9989 - val_loss: 0.3076 - val_categorical_accuracy: 0.8902\n",
            "Epoch 10/30\n",
            "56/56 [==============================] - 10s 169ms/step - loss: 0.0118 - categorical_accuracy: 0.9977 - val_loss: 0.3369 - val_categorical_accuracy: 0.8933\n",
            "Epoch 11/30\n",
            "56/56 [==============================] - 10s 169ms/step - loss: 0.0138 - categorical_accuracy: 0.9955 - val_loss: 0.2897 - val_categorical_accuracy: 0.9055\n",
            "Restoring model weights from the end of the best epoch.\n",
            "Epoch 00011: early stopping\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 216
        },
        "id": "dlOpLVj50cRd",
        "outputId": "28aec72d-7791-4beb-bae7-0c2802f32c98"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "fig, axs = plt.subplots(nrows=1, ncols=2, figsize=(20, 4))\n",
        "\n",
        "axs[0].plot(history.history['loss'], label='loss')\n",
        "axs[0].plot(history.history['val_loss'], label='val_loss')\n",
        "\n",
        "axs[1].plot(history.history['categorical_accuracy'], label='acc')\n",
        "axs[1].plot(history.history['val_categorical_accuracy'], label='val_acc')\n",
        "\n",
        "plt.legend();\n",
        "plt.show();"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MPJbeFaHkyuO",
        "outputId": "3fd733a5-72cd-49dd-e08e-a1e8c7bbdc18"
      },
      "source": [
        "test_generator = datagen.flow_from_dataframe(\n",
        "    dataframe=df_test,\n",
        "    directory=IMAGES_DIR,\n",
        "    x_col='rel_path',\n",
        "    y_col='label',\n",
        "    color_mode=\"rgb\",\n",
        "    class_mode=\"categorical\",\n",
        "    batch_size=1,\n",
        "    shuffle=False,\n",
        "    seed=7\n",
        ")\n",
        "\n",
        "# y_pred = net.predict(test_generator, batch_size=None, verbose=0, steps=None, callbacks=None, max_queue_size=10, workers=1, use_multiprocessing=False)\n",
        "\n",
        "filenames = test_generator.filenames\n",
        "nb_samples = len(filenames)\n",
        "\n",
        "preds = net.predict_generator(test_generator, nb_samples, workers=1)\n",
        "net.evaluate_generator(test_generator, nb_samples)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 328 validated image filenames belonging to 6 classes.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:2001: UserWarning: `Model.predict_generator` is deprecated and will be removed in a future version. Please use `Model.predict`, which supports generators.\n",
            "  warnings.warn('`Model.predict_generator` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:1973: UserWarning: `Model.evaluate_generator` is deprecated and will be removed in a future version. Please use `Model.evaluate`, which supports generators.\n",
            "  warnings.warn('`Model.evaluate_generator` is deprecated and '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.2640801966190338, 0.9024389982223511]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yGc35xBYlnRB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5722ee6b-2096-459d-8afd-105c9fdccfff"
      },
      "source": [
        "Y_test = df_test.copy()\n",
        "Y_test= Y_test.loc[:, 'label'].values\n",
        "print(Y_test)\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['cardboard' 'plastic' 'glass' 'glass' 'glass' 'cardboard' 'metal' 'trash'\n",
            " 'glass' 'paper' 'cardboard' 'metal' 'paper' 'paper' 'glass' 'glass'\n",
            " 'glass' 'paper' 'trash' 'plastic' 'cardboard' 'plastic' 'paper' 'glass'\n",
            " 'plastic' 'glass' 'plastic' 'plastic' 'metal' 'paper' 'paper' 'glass'\n",
            " 'paper' 'cardboard' 'paper' 'glass' 'plastic' 'plastic' 'metal' 'metal'\n",
            " 'metal' 'plastic' 'paper' 'cardboard' 'glass' 'metal' 'metal' 'cardboard'\n",
            " 'glass' 'trash' 'paper' 'paper' 'paper' 'cardboard' 'cardboard' 'plastic'\n",
            " 'glass' 'plastic' 'cardboard' 'metal' 'metal' 'glass' 'metal' 'glass'\n",
            " 'trash' 'glass' 'cardboard' 'glass' 'cardboard' 'paper' 'plastic'\n",
            " 'cardboard' 'trash' 'metal' 'paper' 'paper' 'paper' 'paper' 'glass'\n",
            " 'paper' 'metal' 'glass' 'cardboard' 'plastic' 'metal' 'cardboard' 'glass'\n",
            " 'plastic' 'paper' 'glass' 'paper' 'glass' 'paper' 'plastic' 'cardboard'\n",
            " 'metal' 'paper' 'glass' 'plastic' 'paper' 'paper' 'cardboard' 'glass'\n",
            " 'cardboard' 'plastic' 'plastic' 'glass' 'metal' 'plastic' 'cardboard'\n",
            " 'plastic' 'paper' 'glass' 'plastic' 'cardboard' 'plastic' 'plastic'\n",
            " 'cardboard' 'cardboard' 'metal' 'metal' 'glass' 'cardboard' 'cardboard'\n",
            " 'cardboard' 'paper' 'metal' 'glass' 'paper' 'plastic' 'glass' 'metal'\n",
            " 'plastic' 'glass' 'paper' 'paper' 'plastic' 'trash' 'plastic' 'plastic'\n",
            " 'paper' 'glass' 'plastic' 'paper' 'plastic' 'plastic' 'glass' 'metal'\n",
            " 'glass' 'paper' 'glass' 'glass' 'paper' 'metal' 'plastic' 'metal'\n",
            " 'plastic' 'glass' 'paper' 'glass' 'glass' 'plastic' 'cardboard' 'plastic'\n",
            " 'glass' 'plastic' 'plastic' 'metal' 'cardboard' 'plastic' 'trash' 'glass'\n",
            " 'paper' 'glass' 'plastic' 'glass' 'paper' 'plastic' 'plastic' 'paper'\n",
            " 'glass' 'trash' 'cardboard' 'paper' 'paper' 'cardboard' 'glass' 'paper'\n",
            " 'trash' 'paper' 'metal' 'cardboard' 'paper' 'paper' 'cardboard' 'metal'\n",
            " 'paper' 'metal' 'plastic' 'paper' 'metal' 'paper' 'plastic' 'paper'\n",
            " 'metal' 'cardboard' 'paper' 'glass' 'paper' 'metal' 'paper' 'metal'\n",
            " 'cardboard' 'cardboard' 'paper' 'cardboard' 'plastic' 'metal' 'trash'\n",
            " 'paper' 'paper' 'plastic' 'cardboard' 'glass' 'trash' 'glass' 'paper'\n",
            " 'paper' 'cardboard' 'paper' 'paper' 'paper' 'plastic' 'glass' 'plastic'\n",
            " 'glass' 'trash' 'paper' 'paper' 'plastic' 'cardboard' 'glass' 'glass'\n",
            " 'glass' 'paper' 'glass' 'glass' 'glass' 'cardboard' 'paper' 'metal'\n",
            " 'metal' 'paper' 'glass' 'glass' 'trash' 'paper' 'trash' 'paper' 'paper'\n",
            " 'trash' 'metal' 'cardboard' 'paper' 'paper' 'metal' 'paper' 'cardboard'\n",
            " 'plastic' 'glass' 'metal' 'paper' 'metal' 'plastic' 'plastic' 'metal'\n",
            " 'trash' 'cardboard' 'metal' 'plastic' 'trash' 'paper' 'paper' 'paper'\n",
            " 'metal' 'plastic' 'metal' 'metal' 'paper' 'cardboard' 'metal' 'cardboard'\n",
            " 'glass' 'plastic' 'plastic' 'metal' 'plastic' 'paper' 'metal' 'paper'\n",
            " 'paper' 'metal' 'metal' 'glass' 'metal' 'plastic' 'metal' 'plastic'\n",
            " 'metal' 'metal' 'metal' 'metal' 'paper' 'paper' 'plastic' 'paper' 'glass'\n",
            " 'paper' 'plastic' 'glass' 'metal' 'cardboard' 'glass' 'metal' 'cardboard'\n",
            " 'plastic' 'glass' 'paper']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Cyd68FjJTP7Y",
        "outputId": "933050b3-64e1-41ce-c235-3c900c155c9f"
      },
      "source": [
        "class_dictionary = test_generator.class_indices\n",
        "print(class_dictionary)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'cardboard': 0, 'glass': 1, 'metal': 2, 'paper': 3, 'plastic': 4, 'trash': 5}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QjPXT0OGTSsI"
      },
      "source": [
        "id_class = {'cardboard': 0, 'glass': 1, 'metal': 2, 'paper': 3, 'plastic': 4, 'trash': 5}\n",
        "\n",
        "for id, val in enumerate(Y_test):\n",
        "  Y_test[id] = id_class[val]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YbC0OcItTTxV"
      },
      "source": [
        "Y_preds = list(np.array(np.argmax(preds, axis=1)))\n",
        "Y_test = list(Y_test)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WIZseSHWTU6R",
        "outputId": "6ac20e43-9878-45af-92a0-ce25c0fffb52"
      },
      "source": [
        "from sklearn.metrics import confusion_matrix \n",
        "cmat = confusion_matrix(Y_test, Y_preds) \n",
        "cmat "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[43,  0,  1,  0,  2,  0],\n",
              "       [ 0, 56,  4,  0,  5,  0],\n",
              "       [ 0,  2, 52,  0,  2,  0],\n",
              "       [ 0,  0,  1, 79,  2,  1],\n",
              "       [ 0,  4,  2,  2, 53,  0],\n",
              "       [ 0,  0,  1,  2,  1, 13]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qhRUS8LtTWFL",
        "outputId": "32afa287-c69b-4261-99a2-7231699d58d8"
      },
      "source": [
        "print(\"Top 1 cardboard: \" + str(cmat[0][0]/np.sum(cmat[0])))\n",
        "print(\"Top 1 glass: \" + str(cmat[1][1]/np.sum(cmat[1])))\n",
        "print(\"Top 1 metal: \" + str(cmat[2][2]/np.sum(cmat[2])))\n",
        "print(\"Top 1 paper: \" + str(cmat[3][3]/np.sum(cmat[3])))\n",
        "print(\"Top 1 plastic: \" + str(cmat[4][4]/np.sum(cmat[4])))\n",
        "print(\"Top 1 trash: \" + str(cmat[5][5]/np.sum(cmat[5])))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Top 1 cardboard: 0.9347826086956522\n",
            "Top 1 glass: 0.8615384615384616\n",
            "Top 1 metal: 0.9285714285714286\n",
            "Top 1 paper: 0.9518072289156626\n",
            "Top 1 plastic: 0.8688524590163934\n",
            "Top 1 trash: 0.7647058823529411\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 280
        },
        "id": "PW0o6VqmTXMb",
        "outputId": "c6f61c20-acd2-4e10-ba7e-48152e203d12"
      },
      "source": [
        "import seaborn as sns\n",
        "sns.heatmap(cmat/np.sum(cmat, axis=1), annot=True, fmt='.01%', cmap='Blues')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fce1cd22710>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZjGGlsxQTYTx"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}