transfer_learning_55epochs_with_fine_tuning.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "transfer_learning_55epochs.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "mount_file_id": "1rmG3OVWRRp07djIS-d0lRznrTTTTrr2G",
      "authorship_tag": "ABX9TyNOeCQnjhNevWcnvJcp8eAk",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Nevermetyou65/Overwatch-Heroes-recognition-with-Transfer-learning--Xception--/blob/main/transfer_learning_55epochs_with_fine_tuning.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": "YBfowg11N8ue",
        "outputId": "06fe6de3-6218-4461-ff94-c54fd9693b7a"
      },
      "source": [
        "import os\n",
        "\n",
        "cwd = os.getcwd()\n",
        "print(cwd)\n",
        "data_path = './drive/MyDrive/overwatch_recog/overwatch_newdata/'\n",
        "\n",
        "hero_name = sorted(os.listdir(data_path))\n",
        "print(type(hero_name))\n",
        "print(len(hero_name))\n",
        "hero_dict = {}\n",
        "for i, name in enumerate(hero_name):\n",
        "    hero_dict[i] = name"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content\n",
            "<class 'list'>\n",
            "32\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w5TE-GLmgx-C",
        "outputId": "0277bba3-b39e-44c9-d0bd-c64d77ffdfad"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "tmp = []\n",
        "for sub_folder in os.listdir(data_path):\n",
        "    # print(sub_folder)\n",
        "    images = os.listdir(os.path.join(data_path, sub_folder))\n",
        "    tmp.append(len(images))\n",
        "    print(f\"{sub_folder}: {len(images)}\")\n",
        "\n",
        "print(\"=\"*50)\n",
        "print(\"mean:\", np.mean(tmp))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "widowmaker: 178\n",
            "symmetra: 164\n",
            "roadrog: 132\n",
            "zarya: 153\n",
            "zenyatta: 150\n",
            "winston: 164\n",
            "torbjorn: 145\n",
            "soldier76: 142\n",
            "tracer: 175\n",
            "sombra: 156\n",
            "mercy: 174\n",
            "reaper: 159\n",
            "mei: 154\n",
            "junkrat: 138\n",
            "reinhardt: 153\n",
            "mccree: 157\n",
            "orisa: 136\n",
            "pharah: 151\n",
            "lucio: 137\n",
            "moira: 144\n",
            "baptiste: 105\n",
            "brigitte: 134\n",
            "hanzo: 159\n",
            "ana: 142\n",
            "d.va: 195\n",
            "genji: 156\n",
            "echo: 113\n",
            "bastion: 130\n",
            "doomfist: 114\n",
            "ball: 82\n",
            "ashe&bob: 133\n",
            "sigma: 96\n",
            "==================================================\n",
            "mean: 144.40625\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SsO7KtBkVW1i",
        "outputId": "7292a274-d42d-4684-dcf4-b5f0a816b4b9"
      },
      "source": [
        "print(hero_dict)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{0: 'ana', 1: 'ashe&bob', 2: 'ball', 3: 'baptiste', 4: 'bastion', 5: 'brigitte', 6: 'd.va', 7: 'doomfist', 8: 'echo', 9: 'genji', 10: 'hanzo', 11: 'junkrat', 12: 'lucio', 13: 'mccree', 14: 'mei', 15: 'mercy', 16: 'moira', 17: 'orisa', 18: 'pharah', 19: 'reaper', 20: 'reinhardt', 21: 'roadrog', 22: 'sigma', 23: 'soldier76', 24: 'sombra', 25: 'symmetra', 26: 'torbjorn', 27: 'tracer', 28: 'widowmaker', 29: 'winston', 30: 'zarya', 31: 'zenyatta'}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rtlE7MqEOI7i"
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from PIL import Image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G_VODua9OwOx"
      },
      "source": [
        "# data_gen = ImageDataGenerator(\n",
        "#     rotation_range=20,\n",
        "#     width_shift_range=0.2,\n",
        "#     height_shift_range=0.2,\n",
        "#     shear_range=0.2,\n",
        "#     zoom_range=0.2,\n",
        "#     horizontal_flip=True,\n",
        "#     fill_mode='nearest'\n",
        "# )\n",
        "\n",
        "# test_save_image_path = './drive/MyDrive/overwatch_recog/preview/'\n",
        "# test_image_path = './drive/MyDrive/overwatch_recog/overwatch_heroes/'\n",
        "\n",
        "# test_data = data_gen.flow_from_directory(\n",
        "#     test_image_path,\n",
        "#     target_size=(300, 300),\n",
        "#     batch_size=32,\n",
        "#     save_to_dir=test_save_image_path,\n",
        "#     save_format='jpeg'\n",
        "# )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GB_4v39tTEWy"
      },
      "source": [
        "# (x ,y) = test_data.next()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UgfZjNQzV0pU"
      },
      "source": [
        "# plt.imshow(x[0].astype(np.int32))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-2_ntPPTUdLq"
      },
      "source": [
        "# hero_dict.get(np.argmax(y[0]))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QWRnFoA6W965"
      },
      "source": [
        "#Image pipeline"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pI44dM3MVIRu",
        "outputId": "7c574481-11ca-4a6e-8937-57745bd1aeaa"
      },
      "source": [
        "traindata_gen = ImageDataGenerator(\n",
        "    validation_split=0.2,\n",
        "    rotation_range=15,\n",
        "    width_shift_range=0.2,\n",
        "    height_shift_range=0.2,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.1,\n",
        "    horizontal_flip=True,\n",
        "    fill_mode='nearest',)\n",
        "\n",
        "valdata_gen = ImageDataGenerator(\n",
        "    validation_split=0.2,\n",
        ")\n",
        "# ===========================================================================\n",
        "batch_size = 32\n",
        "traindata_batch = traindata_gen.flow_from_directory(\n",
        "    data_path,\n",
        "    seed=42,\n",
        "    target_size=(300, 300),\n",
        "    batch_size=batch_size,\n",
        "    shuffle=True,\n",
        "    subset='training',\n",
        ")\n",
        "valdata_batch = valdata_gen.flow_from_directory(\n",
        "    data_path,\n",
        "    seed=42,\n",
        "    target_size=(300, 300),\n",
        "    batch_size=batch_size,\n",
        "    subset='validation',\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 3711 images belonging to 32 classes.\n",
            "Found 910 images belonging to 32 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7BPCsDtCZOwm"
      },
      "source": [
        "#Pre-trained model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yGFPOdDNZEAT",
        "outputId": "5e921646-f906-41e5-d32b-406a0165551a"
      },
      "source": [
        "# ลองใช้ Xception\n",
        "base_model = keras.applications.Xception(\n",
        "    weights='imagenet',\n",
        "    input_shape=(300, 300, 3),\n",
        "    include_top=False\n",
        ")\n",
        "\n",
        "base_model.trainable=False"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/xception/xception_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "83689472/83683744 [==============================] - 1s 0us/step\n",
            "83697664/83683744 [==============================] - 1s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jhkYT1jNa37k",
        "outputId": "671f2c82-4c07-4434-f53e-0f455a81faab"
      },
      "source": [
        "# build model\n",
        "inputs = keras.Input(shape=(300,300,3))\n",
        "scaler = layers.Rescaling(scale=1./255)\n",
        "x = scaler(inputs)\n",
        "x = base_model(x, training=False)\n",
        "x = keras.layers.GlobalAveragePooling2D()(x)\n",
        "x = keras.layers.Dropout(0.2)(x)\n",
        "outputs = keras.layers.Dense(units=32, activation='softmax')(x)\n",
        "model = keras.Model(inputs=inputs, outputs=outputs, name='model_with_sigma')\n",
        "\n",
        "optim = keras.optimizers.Adam()\n",
        "loss_fn = keras.losses.CategoricalCrossentropy()\n",
        "model.compile(\n",
        "    optimizer=optim,\n",
        "    loss=loss_fn,\n",
        "    metrics=[keras.metrics.CategoricalAccuracy()]\n",
        ")\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_with_sigma\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_2 (InputLayer)         [(None, 300, 300, 3)]     0         \n",
            "_________________________________________________________________\n",
            "rescaling (Rescaling)        (None, 300, 300, 3)       0         \n",
            "_________________________________________________________________\n",
            "xception (Functional)        (None, 10, 10, 2048)      20861480  \n",
            "_________________________________________________________________\n",
            "global_average_pooling2d (Gl (None, 2048)              0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 2048)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 32)                65568     \n",
            "=================================================================\n",
            "Total params: 20,927,048\n",
            "Trainable params: 65,568\n",
            "Non-trainable params: 20,861,480\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FeYEiHYqg-2A",
        "outputId": "657bd7d5-9454-4653-f118-5657ffbaeaa7"
      },
      "source": [
        "epoch = 55\n",
        "history_1 = model.fit(traindata_batch, epochs=epoch,\n",
        "                      steps_per_epoch=116, \n",
        "                      validation_data=valdata_batch, \n",
        "                      validation_steps=29)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/55\n",
            "116/116 [==============================] - 1560s 13s/step - loss: 3.0495 - categorical_accuracy: 0.2024 - val_loss: 2.3979 - val_categorical_accuracy: 0.4835\n",
            "Epoch 2/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 2.3624 - categorical_accuracy: 0.4034 - val_loss: 1.9161 - val_categorical_accuracy: 0.5802\n",
            "Epoch 3/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 2.0297 - categorical_accuracy: 0.4834 - val_loss: 1.6571 - val_categorical_accuracy: 0.6242\n",
            "Epoch 4/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.8324 - categorical_accuracy: 0.5198 - val_loss: 1.4863 - val_categorical_accuracy: 0.6692\n",
            "Epoch 5/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 1.6932 - categorical_accuracy: 0.5627 - val_loss: 1.3750 - val_categorical_accuracy: 0.6824\n",
            "Epoch 6/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 1.5602 - categorical_accuracy: 0.6009 - val_loss: 1.2962 - val_categorical_accuracy: 0.6956\n",
            "Epoch 7/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.4922 - categorical_accuracy: 0.6004 - val_loss: 1.2248 - val_categorical_accuracy: 0.7110\n",
            "Epoch 8/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.4023 - categorical_accuracy: 0.6284 - val_loss: 1.2086 - val_categorical_accuracy: 0.7044\n",
            "Epoch 9/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.3368 - categorical_accuracy: 0.6492 - val_loss: 1.1369 - val_categorical_accuracy: 0.7308\n",
            "Epoch 10/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.2858 - categorical_accuracy: 0.6707 - val_loss: 1.1003 - val_categorical_accuracy: 0.7396\n",
            "Epoch 11/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.2593 - categorical_accuracy: 0.6669 - val_loss: 1.0854 - val_categorical_accuracy: 0.7374\n",
            "Epoch 12/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.2095 - categorical_accuracy: 0.6761 - val_loss: 1.0638 - val_categorical_accuracy: 0.7297\n",
            "Epoch 13/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.1556 - categorical_accuracy: 0.6931 - val_loss: 1.0289 - val_categorical_accuracy: 0.7462\n",
            "Epoch 14/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 1.1213 - categorical_accuracy: 0.7009 - val_loss: 1.0235 - val_categorical_accuracy: 0.7495\n",
            "Epoch 15/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.0964 - categorical_accuracy: 0.7206 - val_loss: 0.9887 - val_categorical_accuracy: 0.7484\n",
            "Epoch 16/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.0557 - categorical_accuracy: 0.7233 - val_loss: 0.9730 - val_categorical_accuracy: 0.7505\n",
            "Epoch 17/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.0338 - categorical_accuracy: 0.7251 - val_loss: 0.9552 - val_categorical_accuracy: 0.7637\n",
            "Epoch 18/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 1.0163 - categorical_accuracy: 0.7200 - val_loss: 0.9549 - val_categorical_accuracy: 0.7604\n",
            "Epoch 19/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.9714 - categorical_accuracy: 0.7524 - val_loss: 0.9396 - val_categorical_accuracy: 0.7615\n",
            "Epoch 20/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.9576 - categorical_accuracy: 0.7497 - val_loss: 0.9274 - val_categorical_accuracy: 0.7769\n",
            "Epoch 21/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.9533 - categorical_accuracy: 0.7405 - val_loss: 0.9332 - val_categorical_accuracy: 0.7615\n",
            "Epoch 22/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.9228 - categorical_accuracy: 0.7534 - val_loss: 0.9256 - val_categorical_accuracy: 0.7593\n",
            "Epoch 23/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.9075 - categorical_accuracy: 0.7567 - val_loss: 0.9314 - val_categorical_accuracy: 0.7626\n",
            "Epoch 24/55\n",
            "116/116 [==============================] - 174s 2s/step - loss: 0.8809 - categorical_accuracy: 0.7658 - val_loss: 0.9039 - val_categorical_accuracy: 0.7703\n",
            "Epoch 25/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.8607 - categorical_accuracy: 0.7661 - val_loss: 0.9134 - val_categorical_accuracy: 0.7626\n",
            "Epoch 26/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.8457 - categorical_accuracy: 0.7699 - val_loss: 0.8782 - val_categorical_accuracy: 0.7758\n",
            "Epoch 27/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.8427 - categorical_accuracy: 0.7718 - val_loss: 0.8820 - val_categorical_accuracy: 0.7670\n",
            "Epoch 28/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.8216 - categorical_accuracy: 0.7790 - val_loss: 0.8847 - val_categorical_accuracy: 0.7692\n",
            "Epoch 29/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.8029 - categorical_accuracy: 0.7836 - val_loss: 0.8719 - val_categorical_accuracy: 0.7747\n",
            "Epoch 30/55\n",
            "116/116 [==============================] - 174s 2s/step - loss: 0.8097 - categorical_accuracy: 0.7782 - val_loss: 0.8801 - val_categorical_accuracy: 0.7670\n",
            "Epoch 31/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.7946 - categorical_accuracy: 0.7879 - val_loss: 0.8682 - val_categorical_accuracy: 0.7692\n",
            "Epoch 32/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.7780 - categorical_accuracy: 0.7898 - val_loss: 0.8583 - val_categorical_accuracy: 0.7769\n",
            "Epoch 33/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.7648 - categorical_accuracy: 0.7928 - val_loss: 0.8533 - val_categorical_accuracy: 0.7813\n",
            "Epoch 34/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.7281 - categorical_accuracy: 0.8041 - val_loss: 0.8542 - val_categorical_accuracy: 0.7824\n",
            "Epoch 35/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.7342 - categorical_accuracy: 0.7984 - val_loss: 0.8597 - val_categorical_accuracy: 0.7835\n",
            "Epoch 36/55\n",
            "116/116 [==============================] - 174s 2s/step - loss: 0.7230 - categorical_accuracy: 0.8057 - val_loss: 0.8487 - val_categorical_accuracy: 0.7736\n",
            "Epoch 37/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.7072 - categorical_accuracy: 0.8100 - val_loss: 0.8457 - val_categorical_accuracy: 0.7769\n",
            "Epoch 38/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.7060 - categorical_accuracy: 0.8149 - val_loss: 0.8491 - val_categorical_accuracy: 0.7692\n",
            "Epoch 39/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.7169 - categorical_accuracy: 0.8006 - val_loss: 0.8467 - val_categorical_accuracy: 0.7725\n",
            "Epoch 40/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.6873 - categorical_accuracy: 0.8162 - val_loss: 0.8382 - val_categorical_accuracy: 0.7769\n",
            "Epoch 41/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.6895 - categorical_accuracy: 0.8081 - val_loss: 0.8389 - val_categorical_accuracy: 0.7714\n",
            "Epoch 42/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.6699 - categorical_accuracy: 0.8135 - val_loss: 0.8461 - val_categorical_accuracy: 0.7670\n",
            "Epoch 43/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.6688 - categorical_accuracy: 0.8176 - val_loss: 0.8431 - val_categorical_accuracy: 0.7791\n",
            "Epoch 44/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.6537 - categorical_accuracy: 0.8178 - val_loss: 0.8183 - val_categorical_accuracy: 0.7945\n",
            "Epoch 45/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.6675 - categorical_accuracy: 0.8130 - val_loss: 0.8436 - val_categorical_accuracy: 0.7802\n",
            "Epoch 46/55\n",
            "116/116 [==============================] - 175s 2s/step - loss: 0.6336 - categorical_accuracy: 0.8257 - val_loss: 0.8251 - val_categorical_accuracy: 0.7813\n",
            "Epoch 47/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.6433 - categorical_accuracy: 0.8238 - val_loss: 0.8288 - val_categorical_accuracy: 0.7802\n",
            "Epoch 48/55\n",
            "116/116 [==============================] - 174s 2s/step - loss: 0.6234 - categorical_accuracy: 0.8316 - val_loss: 0.8134 - val_categorical_accuracy: 0.7846\n",
            "Epoch 49/55\n",
            "116/116 [==============================] - 175s 2s/step - loss: 0.6301 - categorical_accuracy: 0.8235 - val_loss: 0.8226 - val_categorical_accuracy: 0.7824\n",
            "Epoch 50/55\n",
            "116/116 [==============================] - 173s 1s/step - loss: 0.6154 - categorical_accuracy: 0.8302 - val_loss: 0.8313 - val_categorical_accuracy: 0.7868\n",
            "Epoch 51/55\n",
            "116/116 [==============================] - 175s 2s/step - loss: 0.6334 - categorical_accuracy: 0.8305 - val_loss: 0.8314 - val_categorical_accuracy: 0.7879\n",
            "Epoch 52/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.6173 - categorical_accuracy: 0.8292 - val_loss: 0.8293 - val_categorical_accuracy: 0.7725\n",
            "Epoch 53/55\n",
            "116/116 [==============================] - 175s 2s/step - loss: 0.5841 - categorical_accuracy: 0.8405 - val_loss: 0.8146 - val_categorical_accuracy: 0.7857\n",
            "Epoch 54/55\n",
            "116/116 [==============================] - 174s 2s/step - loss: 0.6005 - categorical_accuracy: 0.8375 - val_loss: 0.8119 - val_categorical_accuracy: 0.7824\n",
            "Epoch 55/55\n",
            "116/116 [==============================] - 174s 1s/step - loss: 0.5865 - categorical_accuracy: 0.8391 - val_loss: 0.8292 - val_categorical_accuracy: 0.7747\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F8q6oUhqhr75",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b2010111-56fa-4242-aeb8-a85aed560337"
      },
      "source": [
        "save_path = './drive/MyDrive/overwatch_recog/55epoch_xception_newdata/'\n",
        "model.save(save_path)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: ./drive/MyDrive/overwatch_recog/55epoch_xception_newdata/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/utils/generic_utils.py:497: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
            "  category=CustomMaskWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EJsbvrqGgwWj"
      },
      "source": [
        "# history_1.history.keys()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m72VWFWlhOhv"
      },
      "source": [
        "def plot_history(epoch, history):\n",
        "    epoch = np.arange(epoch)\n",
        "    fig = plt.figure(figsize=(20, 12))\n",
        "    ax = fig.add_subplot(1, 2, 1)\n",
        "    ax.plot(epoch, history.history.get('loss'), label='train loss')\n",
        "    ax.plot(epoch, history.history.get('val_loss'), label='val losss')\n",
        "    plt.legend()\n",
        "    ax2 = fig.add_subplot(1, 2, 2)\n",
        "    ax2.plot(epoch, history.history.get('categorical_accuracy'), label='train acc')\n",
        "    ax2.plot(epoch, history.history.get('val_categorical_accuracy'), label='val acc')\n",
        "    plt.legend()\n",
        "    plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 700
        },
        "id": "kv7-o0z5hZk6",
        "outputId": "a288a756-f921-43cf-8d5d-95506f841e44"
      },
      "source": [
        "plot_history(55, history_1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x864 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2QCifISYhMrS"
      },
      "source": [
        "# Fine-tuning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ImN7sjZShFAO"
      },
      "source": [
        "base_model.trainable = True\n",
        "\n",
        "optim = keras.optimizers.Adam(1e-5)\n",
        "loss_fn = keras.losses.CategoricalCrossentropy()\n",
        "model.compile(\n",
        "    optimizer=optim,\n",
        "    loss=loss_fn,\n",
        "    metrics=[keras.metrics.CategoricalAccuracy()]\n",
        ")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FE_-lACUhmw3",
        "outputId": "d9a7b179-fae2-4643-f2e3-bd160f9ff0b6"
      },
      "source": [
        "epoch = 10\n",
        "history_2 = model.fit(traindata_batch, epochs=epoch,\n",
        "                      steps_per_epoch=116, \n",
        "                      validation_data=valdata_batch, \n",
        "                      validation_steps=29)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "116/116 [==============================] - 184s 2s/step - loss: 0.4486 - categorical_accuracy: 0.8650 - val_loss: 0.6457 - val_categorical_accuracy: 0.8242\n",
            "Epoch 2/10\n",
            "116/116 [==============================] - 177s 2s/step - loss: 0.3426 - categorical_accuracy: 0.8957 - val_loss: 0.6103 - val_categorical_accuracy: 0.8396\n",
            "Epoch 3/10\n",
            "116/116 [==============================] - 177s 2s/step - loss: 0.2958 - categorical_accuracy: 0.9100 - val_loss: 0.6369 - val_categorical_accuracy: 0.8297\n",
            "Epoch 4/10\n",
            "116/116 [==============================] - 176s 2s/step - loss: 0.2324 - categorical_accuracy: 0.9256 - val_loss: 0.5958 - val_categorical_accuracy: 0.8615\n",
            "Epoch 5/10\n",
            "116/116 [==============================] - 176s 2s/step - loss: 0.2244 - categorical_accuracy: 0.9264 - val_loss: 0.5591 - val_categorical_accuracy: 0.8637\n",
            "Epoch 6/10\n",
            "116/116 [==============================] - 177s 2s/step - loss: 0.1986 - categorical_accuracy: 0.9394 - val_loss: 0.5834 - val_categorical_accuracy: 0.8571\n",
            "Epoch 7/10\n",
            "116/116 [==============================] - 177s 2s/step - loss: 0.1547 - categorical_accuracy: 0.9512 - val_loss: 0.5716 - val_categorical_accuracy: 0.8571\n",
            "Epoch 8/10\n",
            "116/116 [==============================] - 177s 2s/step - loss: 0.1437 - categorical_accuracy: 0.9569 - val_loss: 0.5726 - val_categorical_accuracy: 0.8670\n",
            "Epoch 9/10\n",
            "116/116 [==============================] - 177s 2s/step - loss: 0.1281 - categorical_accuracy: 0.9601 - val_loss: 0.5743 - val_categorical_accuracy: 0.8637\n",
            "Epoch 10/10\n",
            "116/116 [==============================] - 176s 2s/step - loss: 0.1164 - categorical_accuracy: 0.9625 - val_loss: 0.5561 - val_categorical_accuracy: 0.8593\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 700
        },
        "id": "EWRQMos0hzt2",
        "outputId": "6910b5a0-fc59-48ca-bc03-ab17094c00d7"
      },
      "source": [
        "plot_history(10, history_2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x864 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZKmJTaQEorBI",
        "outputId": "af4b3a4d-ae69-4bcb-de57-7b21e2cbf653"
      },
      "source": [
        "save_path = './drive/MyDrive/overwatch_recog/65epoch_xception_fine_tuning/'\n",
        "model.save(save_path)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: ./drive/MyDrive/overwatch_recog/65epoch_xception_fine_tuning/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/utils/generic_utils.py:497: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
            "  category=CustomMaskWarning)\n"
          ]
        }
      ]
    }
  ]
}