LAB5_Basic_CNN.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y4bJXWxZJ4s0"
      },
      "source": [
        "## Convolution neural network (CNN)\n",
        "\n",
        "โหลดชุดข้อมูลภาพ Cifar-10 โดยใช้ keras"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "m8bcy-lhJ4s4",
        "outputId": "c1c821ce-2ee8-42ea-a931-b182df48238f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
            "170500096/170498071 [==============================] - 2s 0us/step\n",
            "170508288/170498071 [==============================] - 2s 0us/step\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "from tensorflow.keras.datasets import cifar10\n",
        "from keras.utils.np_utils import to_categorical     # convert label to binary vector\n",
        "\n",
        "(X_train, y_train), (X_test, y_test) = cifar10.load_data()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7upO2T0AMJcR"
      },
      "source": [
        "แสดงมิติชุดข้อมูลภาพด้วย ชุดข้อมูลฝึกฝน (train data) และ ชุดข้อมูลทดสอบ (test data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QNS94VQMJ4s_",
        "outputId": "be7e0031-432e-4af1-c17f-4c159b77338a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Shape of training data:\n",
            "(50000, 32, 32, 3)\n",
            "(50000, 1)\n",
            "Shape of test data:\n",
            "(10000, 32, 32, 3)\n",
            "(10000, 1)\n"
          ]
        }
      ],
      "source": [
        "print(\"Shape of training data:\")\n",
        "print(X_train.shape)\n",
        "print(y_train.shape)\n",
        "print(\"Shape of test data:\")\n",
        "print(X_test.shape)\n",
        "print(y_test.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WXOl0DAdJn77"
      },
      "source": [
        "แสดงภาพข้อมูลจากชุดข้อมูลฝุกฝน (train data) ด้วย matplotlib"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        },
        "id": "5j9Y61SbJ4tE",
        "outputId": "3a754e89-bbb7-48f9-f167-cfddc8a2b4c0"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Example training images and their labels: [6, 9, 9, 4, 1]\n",
            "Corresponding classes for the labels: ['frog', 'truck', 'truck', 'deer', 'automobile']\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x432 with 5 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "cifar_classes = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']\n",
        "print('Example training images and their labels: ' + str([x[0] for x in y_train[0:5]]))\n",
        "print('Corresponding classes for the labels: ' + str([cifar_classes[x[0]] for x in y_train[0:5]]))\n",
        "\n",
        "f, axarr = plt.subplots(1, 5)\n",
        "f.set_size_inches(16, 6)\n",
        "\n",
        "for i in range(5):\n",
        "    img = X_train[i]\n",
        "    axarr[i].imshow(img)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1wWTIcTSJz5A"
      },
      "source": [
        "เข้ารหัสลาเบล (label) ให้อยู่ในรูป binary code vector สำหรับเอาท์ุพุตของโมเดล cnn\n",
        "\n",
        "นอมัลไลน์ข้อมูลภาพ  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ags3pPvZJ4tJ"
      },
      "outputs": [],
      "source": [
        "y_train = to_categorical(y_train)\n",
        "y_test = to_categorical(y_test)\n",
        "\n",
        "X_train = X_train.astype('float32')\n",
        "X_test = X_test.astype('float32')\n",
        "X_train /= 255\n",
        "X_test /= 255"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DD6wH-C2KhTR"
      },
      "source": [
        "สร้างโมเดล CNN เพื่อจำแนกภาพ"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "boF98qLQJ4tR",
        "outputId": "7ddc3255-7205-4df4-fa30-3600aa010f94"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 30, 30, 32)        896       \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 15, 15, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 13, 13, 32)        9248      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 6, 6, 32)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 1152)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 10)                11530     \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                110       \n",
            "=================================================================\n",
            "Total params: 21,784\n",
            "Trainable params: 21,784\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.layers import Dense, Flatten\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.optimizers import SGD\n",
        "\n",
        "cnn = Sequential()\n",
        "cnn.add(Conv2D(32, (3, 3), activation='relu', input_shape=(32, 32, 3)))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "cnn.add(Conv2D(32, (3, 3), activation='relu'))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "cnn.add(Flatten())\n",
        "cnn.add(Dense(10, activation='relu'))\n",
        "cnn.add(Dense(10, activation='softmax'))\n",
        "\n",
        "sgd = SGD(learning_rate=0.01)\n",
        "\n",
        "cnn.compile(loss='categorical_crossentropy', metrics=['accuracy'], optimizer=sgd)\n",
        "print(cnn.summary())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CvRbG7AKKwsO"
      },
      "source": [
        "ฝึกฝนโมเดล cnn ที่ถูกสร้างด้านบนบนชุดข้อมูลฝึกฝน (train data) และ plot กราฟแสดงค่า loss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "w41IJsDSr27O"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jLPulax1NTN2"
      },
      "source": [
        "แสดงค่า loss และ accuracy ของโมเดลที่ถูกฝึกฝนแล้ว บนชุดข้อมูล test"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-u_ci2nqJ4ta",
        "outputId": "ba6dbed2-9d0d-4376-dedf-1b69a28e11b3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "['loss', 'accuracy']\n",
            "[1.1491912603378296, 0.6001999974250793]\n"
          ]
        }
      ],
      "source": [
        "score = cnn.evaluate(X_test, y_test, verbose=0)\n",
        "# y= cnn.predict(img)\n",
        "print(cnn.metrics_names)\n",
        "print(score)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jLJPQiu4J4te"
      },
      "source": [
        "# Exercise\n",
        "\n",
        "## Build Convolution Neural Network for cat&dog image data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EqzyopQSoqLP",
        "outputId": "6beea7d7-d78c-4fbb-c12b-ddeb4d4716e9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Cloning into 'Cats-Dogs-with-keras'...\n",
            "remote: Enumerating objects: 2893, done.\u001b[K\n",
            "remote: Counting objects: 100% (2893/2893), done.\u001b[K\n",
            "remote: Compressing objects: 100% (2892/2892), done.\u001b[K\n",
            "remote: Total 2893 (delta 0), reused 2893 (delta 0), pack-reused 0\u001b[K\n",
            "Receiving objects: 100% (2893/2893), 62.53 MiB | 18.98 MiB/s, done.\n"
          ]
        }
      ],
      "source": [
        "!git clone https://github.com/vyomshm/Cats-Dogs-with-keras.git"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gVVnoQjho35z",
        "outputId": "b1b12c2b-424e-4597-9446-1d1cc5da9ffe"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(396, 312, 3)\n"
          ]
        }
      ],
      "source": [
        "import cv2\n",
        "img = cv2.imread('Cats-Dogs-with-keras/data/train/cats/cat.2.jpg')\n",
        "print(img.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "11pMG2TZora8"
      },
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FOHkHrsaJ4th",
        "outputId": "83abe18f-e2bc-4bc8-da12-cff51930a4c3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Reading images from path: Cats-Dogs-with-keras/data/train/cats\n",
            "Reading images from path: Cats-Dogs-with-keras/data/train/dogs\n",
            "(1638, 32, 32, 3)\n",
            "(1638, 2)\n"
          ]
        }
      ],
      "source": [
        "import cv2\n",
        "import os                                         ## package for file path system\n",
        "import numpy as np\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras.utils import to_categorical            ## encode numbers to one-hot encode form\n",
        "from sklearn.preprocessing import LabelEncoder    ## encode string labels to numbers\n",
        "\n",
        "folders = ['Cats-Dogs-with-keras/data/train/cats',      #define paths for image files\n",
        "           'Cats-Dogs-with-keras/data/train/dogs'\n",
        "          ]\n",
        "\n",
        "images = []                                                   # input image list\n",
        "labels = []                                                   # image label list\n",
        "\n",
        "def load_images_from_folder(folder):             # function to read image files from a given folder\n",
        "    for filename in os.listdir(folder):          # for each image file in the folder\n",
        "        img = cv2.imread(os.path.join(folder, filename))  # read an image file\n",
        "        img = cv2.resize(img, (32, 32))\n",
        "        #img =image.img_to_array(img)          # convert image data to keras array\n",
        "        images.append(img)                         # append the image data\n",
        "        label = folder.split(os.path.sep)[-1]      # get a label for the image\n",
        "        labels.append(label)                       # append the label\n",
        "\n",
        "def convert_images(images):\n",
        "    X = np.array(images)\n",
        "    # X = X.reshape(X.shape[0],100*100*3)            ## reshape 3D to 1D vector\n",
        "    return X\n",
        "\n",
        "def convert_labels(labels):\n",
        "    enc = LabelEncoder()\n",
        "    P = enc.fit_transform(labels)                ## convert string labels to numbers\n",
        "    Y = to_categorical(P)                        ## convert number to one-hot-encode form\n",
        "    return Y\n",
        "\n",
        "for folder in folders:\n",
        "    print('Reading images from path: ' + folder)\n",
        "    load_images_from_folder(folder)\n",
        "X = np.array(images)/255.0\n",
        "Y = convert_labels(labels)\n",
        "\n",
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = 0.2)\n",
        "\n",
        "print(X_train.shape)\n",
        "print(Y_train.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qVc3LSbCiyBy"
      },
      "source": [
        "## ออกแบบ CNN for Cat & Dog classification"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Dmm05qkZpUEr"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Activation, Flatten\n",
        "from tensorflow.keras.optimizers import SGD\n",
        "\n",
        "cnn = Sequential()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Yq4BFAVti_a5"
      },
      "source": [
        "### ฝึกฝนโมเดล CNN ด้วยชุดข้อมูลฝึกฝน และ plot กราฟแสดง loss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QDlQY9HxBUN7"
      },
      "outputs": [],
      "source": [
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NFzrF7h3Iotp"
      },
      "source": [
        "วัดผลการฝึกฝนโมเดล cnn บนชุดข้อมูลทดสอบ (test data) โดยแสดงค่า loss และ accuracy"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fgcsH7gVEfaI",
        "outputId": "9a539e23-5d07-471f-a3e3-9aa18a3c25f4"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "13/13 [==============================] - 0s 10ms/step - loss: 1.0405 - accuracy: 0.5415\n",
            "[1.040503740310669, 0.5414634346961975]\n"
          ]
        }
      ],
      "source": [
        "score = cnn.evaluate(X_test,Y_test)\n",
        "print(score)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Es8XHh3YA-Lv"
      },
      "source": [
        "โหลดรูปภาพสุนัขหรือแมวจากภายนอก จากนั้นสั่งให้ CNN ทำการทำนายลาเบล"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 304
        },
        "id": "FR9AbFNfA90S",
        "outputId": "e5ba8ba9-70ff-442a-c5d5-280e1f9fd89b"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[[0. 1.]]\n",
            "label :  dog\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "labels = ['cat', 'dog']\n",
        "new_img = cv2.imread('test1.jpg')\n",
        "plt.imshow(new_img)\n",
        "\n",
        "new_img = cv2.resize(new_img, (32, 32))\n",
        "\n",
        "new_img = new_img.reshape(1, 32, 32, 3)\n",
        "\n",
        "predicted_label = cnn.predict(new_img)\n",
        "print(predicted_label)\n",
        "print('label : ', labels[np.argmax(predicted_label[0])])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "orBcdkz0BOWZ"
      },
      "source": [
        "สร้างโมเดล CNN LeNet-5 โดยมีโครงสร้างตามด้านล่าง จากนั้นทดลองฝึกฝนโมเดลเพื่อจำแนก dog & cat\n",
        "\n",
        "![](https://miro.medium.com/max/4000/1*DMcPgeekUftwk0GTMcNawg.png)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TGf3jhxrBN8S"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c3Iphh-RjLhN"
      },
      "source": [
        "## Keras Applications"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FN_KwmY41oOM",
        "outputId": "d5847b97-099b-4446-add6-b095325638bf"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels.h5\n",
            "102973440/102967424 [==============================] - 2s 0us/step\n",
            "102981632/102967424 [==============================] - 2s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/download.tensorflow.org/data/imagenet_class_index.json\n",
            "40960/35363 [==================================] - 0s 0us/step\n",
            "49152/35363 [=========================================] - 0s 0us/step\n",
            "Predicted: [('n02110806', 'basenji', 0.6543896), ('n02091244', 'Ibizan_hound', 0.1471521), ('n02093428', 'American_Staffordshire_terrier', 0.09510354)]\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.applications.resnet50 import ResNet50\n",
        "from tensorflow.keras.preprocessing import image\n",
        "from tensorflow.keras.applications.resnet50 import preprocess_input, decode_predictions\n",
        "import numpy as np\n",
        "\n",
        "model = ResNet50(weights='imagenet')\n",
        "\n",
        "img_path = 'test1.jpg'\n",
        "img = image.load_img(img_path, target_size=(224, 224))\n",
        "x = image.img_to_array(img)\n",
        "x = np.expand_dims(x, axis=0)\n",
        "x = preprocess_input(x)\n",
        "\n",
        "preds = model.predict(x)\n",
        "# decode the results into a list of tuples (class, description, probability)\n",
        "# (one such list for each sample in the batch)\n",
        "print('Predicted:', decode_predictions(preds, top=3)[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9uZebpOWTlHN"
      },
      "source": [
        "# Transfer learning\n",
        "Transfer learning consists of taking features learned on one problem, and leveraging them on a new, similar problem. It is usually done for tasks where your dataset has too little data to train a full-scale model from scratch.\n",
        "![](https://www.researchgate.net/publication/328680040/figure/fig1/AS:688421848752129@1541143967564/Schematic-representation-of-convolutional-neural-network-CNN-architecture-and-the.png)\n",
        "\n",
        "\n",
        "ศึกษาการทำ transfer learning ใน keras จากลิงค์ด้านล่าง\n",
        "https://machinelearningmastery.com/how-to-use-transfer-learning-when-developing-convolutional-neural-network-models/\n",
        "\n",
        "จากนั้นทดลองประยุกต์ใช้ pretrained VGG16 สำหรับการจำแนก dog & cat และ cifar10\n",
        "![](https://storage.googleapis.com/lds-media/images/vgg16-architecture.width-1200.jpg)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "skG78in4KBXB",
        "outputId": "bf1fcfd9-56c3-4d8c-e867-7cba61397826"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"vgg16\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_4 (InputLayer)         [(None, 32, 32, 3)]       0         \n",
            "_________________________________________________________________\n",
            "block1_conv1 (Conv2D)        (None, 32, 32, 64)        1792      \n",
            "_________________________________________________________________\n",
            "block1_conv2 (Conv2D)        (None, 32, 32, 64)        36928     \n",
            "_________________________________________________________________\n",
            "block1_pool (MaxPooling2D)   (None, 16, 16, 64)        0         \n",
            "_________________________________________________________________\n",
            "block2_conv1 (Conv2D)        (None, 16, 16, 128)       73856     \n",
            "_________________________________________________________________\n",
            "block2_conv2 (Conv2D)        (None, 16, 16, 128)       147584    \n",
            "_________________________________________________________________\n",
            "block2_pool (MaxPooling2D)   (None, 8, 8, 128)         0         \n",
            "_________________________________________________________________\n",
            "block3_conv1 (Conv2D)        (None, 8, 8, 256)         295168    \n",
            "_________________________________________________________________\n",
            "block3_conv2 (Conv2D)        (None, 8, 8, 256)         590080    \n",
            "_________________________________________________________________\n",
            "block3_conv3 (Conv2D)        (None, 8, 8, 256)         590080    \n",
            "_________________________________________________________________\n",
            "block3_pool (MaxPooling2D)   (None, 4, 4, 256)         0         \n",
            "_________________________________________________________________\n",
            "block4_conv1 (Conv2D)        (None, 4, 4, 512)         1180160   \n",
            "_________________________________________________________________\n",
            "block4_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block4_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block4_pool (MaxPooling2D)   (None, 2, 2, 512)         0         \n",
            "_________________________________________________________________\n",
            "block5_conv1 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv2 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv3 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block5_pool (MaxPooling2D)   (None, 1, 1, 512)         0         \n",
            "=================================================================\n",
            "Total params: 14,714,688\n",
            "Trainable params: 0\n",
            "Non-trainable params: 14,714,688\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.applications import VGG16\n",
        "\n",
        "base_model = VGG16(weights='imagenet',  # Load weights pre-trained on ImageNet.\n",
        "                   input_shape=(32, 32, 3),\n",
        "                    include_top=False)  # Do not include the ImageNet classifier at the top.\n",
        "\n",
        "base_model.trainable = False\n",
        "\n",
        "base_model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7GzNv_TN5O5H",
        "outputId": "51fb832b-02d6-445b-c65d-fb55eda05566"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_5 (InputLayer)         [(None, 32, 32, 3)]       0         \n",
            "_________________________________________________________________\n",
            "vgg16 (Functional)           (None, 1, 1, 512)         14714688  \n",
            "_________________________________________________________________\n",
            "global_average_pooling2d (Gl (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_12 (Dense)             (None, 2)                 1026      \n",
            "=================================================================\n",
            "Total params: 14,715,714\n",
            "Trainable params: 1,026\n",
            "Non-trainable params: 14,714,688\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "from tensorflow import keras\n",
        "\n",
        "inputs = keras.Input(shape=(32, 32, 3))\n",
        "x = base_model(inputs, training=False)\n",
        "x = keras.layers.GlobalAveragePooling2D()(x)\n",
        "outputs = keras.layers.Dense(2, activation='softmax')(x)\n",
        "\n",
        "model = keras.Model(inputs, outputs)\n",
        "model.summary()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jFjt0gcT65_M",
        "outputId": "cdb1ee45-8d28-42a7-afea-a89963ba6255"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 24ms/step - loss: 0.7752 - accuracy: 0.5324\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6874 - accuracy: 0.6093\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6816 - accuracy: 0.6215\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6845 - accuracy: 0.6074\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 19ms/step - loss: 0.6514 - accuracy: 0.6416\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6229 - accuracy: 0.6587\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6568 - accuracy: 0.6477\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.5978 - accuracy: 0.6752\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 19ms/step - loss: 0.5919 - accuracy: 0.6844\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6084 - accuracy: 0.6728\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6159 - accuracy: 0.6606\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6274 - accuracy: 0.6691\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.5838 - accuracy: 0.7009\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.5687 - accuracy: 0.6990\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 0.6070 - accuracy: 0.6783\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f9f36112650>"
            ]
          },
          "execution_count": 35,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "sgd = SGD(learning_rate=0.03)\n",
        "model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])\n",
        "\n",
        "model.fit(X_train, Y_train, epochs=15)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B2acMIF2J4tl"
      },
      "source": [
        "## Homework\n",
        "\n",
        "### โหลดชุดภาพ hand gesture image dataset จากลิงค์ด้านล่าง\n",
        "### https://github.com/ardamavi/Sign-Language-Digits-Dataset/tree/master/Dataset\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QeSrK5RXJ4tm",
        "outputId": "47cd1c9d-b421-492d-8b87-22f38bff1c5e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Cloning into 'Sign-Language-Digits-Dataset'...\n",
            "remote: Enumerating objects: 2095, done.\u001b[K\n",
            "remote: Counting objects: 100% (6/6), done.\u001b[K\n",
            "remote: Compressing objects: 100% (6/6), done.\u001b[K\n",
            "remote: Total 2095 (delta 2), reused 0 (delta 0), pack-reused 2089\u001b[K\n",
            "Receiving objects: 100% (2095/2095), 15.07 MiB | 24.34 MiB/s, done.\n",
            "Resolving deltas: 100% (660/660), done.\n"
          ]
        }
      ],
      "source": [
        "!git clone https://github.com/ardamavi/Sign-Language-Digits-Dataset.git"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "R_D-hm6ifXYV",
        "outputId": "be192194-9aab-4cd6-d68e-fe4d6b60509d"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/0\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/1\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/2\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/3\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/4\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/5\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/6\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/7\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/8\n",
            "Reading images from path: Sign-Language-Digits-Dataset/Dataset/9\n",
            "training data :  (1649, 32, 32, 3)\n",
            "test data :  (413, 32, 32, 3)\n",
            "Model: \"sequential_10\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_40 (Conv2D)           (None, 30, 30, 32)        896       \n",
            "_________________________________________________________________\n",
            "conv2d_41 (Conv2D)           (None, 28, 28, 32)        9248      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_20 (MaxPooling (None, 14, 14, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_42 (Conv2D)           (None, 12, 12, 64)        18496     \n",
            "_________________________________________________________________\n",
            "conv2d_43 (Conv2D)           (None, 10, 10, 64)        36928     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_21 (MaxPooling (None, 5, 5, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten_10 (Flatten)         (None, 1600)              0         \n",
            "_________________________________________________________________\n",
            "dense_20 (Dense)             (None, 50)                80050     \n",
            "_________________________________________________________________\n",
            "dense_21 (Dense)             (None, 10)                510       \n",
            "=================================================================\n",
            "Total params: 146,128\n",
            "Trainable params: 146,128\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Epoch 1/100\n",
            "26/26 [==============================] - 1s 23ms/step - loss: 2.3030 - val_loss: 2.2999\n",
            "Epoch 2/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 2.2976 - val_loss: 2.2942\n",
            "Epoch 3/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.2860 - val_loss: 2.2776\n",
            "Epoch 4/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 2.2409 - val_loss: 2.2093\n",
            "Epoch 5/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 2.1203 - val_loss: 1.7463\n",
            "Epoch 6/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.0560 - val_loss: 1.8007\n",
            "Epoch 7/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.3949 - val_loss: 0.9460\n",
            "Epoch 8/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.9085 - val_loss: 0.6635\n",
            "Epoch 9/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 0.5795 - val_loss: 0.4906\n",
            "Epoch 10/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.4007 - val_loss: 0.3823\n",
            "Epoch 11/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.3566 - val_loss: 0.3661\n",
            "Epoch 12/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.2713 - val_loss: 0.2762\n",
            "Epoch 13/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.3442 - val_loss: 0.2531\n",
            "Epoch 14/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 0.1583 - val_loss: 0.2515\n",
            "Epoch 15/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.1218 - val_loss: 0.2577\n",
            "Epoch 16/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 0.0962 - val_loss: 0.2364\n",
            "Epoch 17/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0794 - val_loss: 0.3095\n",
            "Epoch 18/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 0.0610 - val_loss: 0.2433\n",
            "Epoch 19/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0435 - val_loss: 0.3837\n",
            "Epoch 20/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0464 - val_loss: 0.2339\n",
            "Epoch 21/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 0.0136 - val_loss: 0.3141\n",
            "Epoch 22/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 0.0169 - val_loss: 0.2733\n",
            "Epoch 23/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0057 - val_loss: 0.2850\n",
            "Epoch 24/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 0.0035 - val_loss: 0.2818\n",
            "Epoch 25/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0026 - val_loss: 0.2929\n",
            "Epoch 26/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0021 - val_loss: 0.2943\n",
            "Epoch 27/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 0.0017 - val_loss: 0.3042\n",
            "Epoch 28/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0015 - val_loss: 0.3066\n",
            "Epoch 29/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 0.0013 - val_loss: 0.3119\n",
            "Epoch 30/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 0.0011 - val_loss: 0.3176\n",
            "Epoch 31/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 0.0010 - val_loss: 0.3180\n",
            "Epoch 32/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 9.3100e-04 - val_loss: 0.3229\n",
            "Epoch 33/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 8.3269e-04 - val_loss: 0.3268\n",
            "Epoch 34/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 7.6692e-04 - val_loss: 0.3274\n",
            "Epoch 35/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 7.1118e-04 - val_loss: 0.3307\n",
            "Epoch 36/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 6.6226e-04 - val_loss: 0.3353\n",
            "Epoch 37/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 6.1676e-04 - val_loss: 0.3369\n",
            "Epoch 38/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 5.7073e-04 - val_loss: 0.3383\n",
            "Epoch 39/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 5.3850e-04 - val_loss: 0.3415\n",
            "Epoch 40/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 5.0805e-04 - val_loss: 0.3427\n",
            "Epoch 41/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 4.8186e-04 - val_loss: 0.3427\n",
            "Epoch 42/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 4.5480e-04 - val_loss: 0.3451\n",
            "Epoch 43/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 4.3002e-04 - val_loss: 0.3475\n",
            "Epoch 44/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 4.1205e-04 - val_loss: 0.3500\n",
            "Epoch 45/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 3.9393e-04 - val_loss: 0.3515\n",
            "Epoch 46/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 3.7600e-04 - val_loss: 0.3531\n",
            "Epoch 47/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 3.5844e-04 - val_loss: 0.3546\n",
            "Epoch 48/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 3.4402e-04 - val_loss: 0.3564\n",
            "Epoch 49/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 3.2800e-04 - val_loss: 0.3566\n",
            "Epoch 50/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 3.1767e-04 - val_loss: 0.3587\n",
            "Epoch 51/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 3.0531e-04 - val_loss: 0.3600\n",
            "Epoch 52/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.9449e-04 - val_loss: 0.3618\n",
            "Epoch 53/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.8425e-04 - val_loss: 0.3626\n",
            "Epoch 54/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 2.7411e-04 - val_loss: 0.3638\n",
            "Epoch 55/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.6489e-04 - val_loss: 0.3654\n",
            "Epoch 56/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.5595e-04 - val_loss: 0.3665\n",
            "Epoch 57/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 2.4835e-04 - val_loss: 0.3673\n",
            "Epoch 58/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.4009e-04 - val_loss: 0.3690\n",
            "Epoch 59/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.3306e-04 - val_loss: 0.3700\n",
            "Epoch 60/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.2531e-04 - val_loss: 0.3709\n",
            "Epoch 61/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.1909e-04 - val_loss: 0.3715\n",
            "Epoch 62/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.1318e-04 - val_loss: 0.3727\n",
            "Epoch 63/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 2.0750e-04 - val_loss: 0.3735\n",
            "Epoch 64/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 2.0213e-04 - val_loss: 0.3746\n",
            "Epoch 65/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.9665e-04 - val_loss: 0.3753\n",
            "Epoch 66/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.9129e-04 - val_loss: 0.3763\n",
            "Epoch 67/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.8663e-04 - val_loss: 0.3769\n",
            "Epoch 68/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.8186e-04 - val_loss: 0.3771\n",
            "Epoch 69/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.7689e-04 - val_loss: 0.3793\n",
            "Epoch 70/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.7346e-04 - val_loss: 0.3795\n",
            "Epoch 71/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.6876e-04 - val_loss: 0.3795\n",
            "Epoch 72/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.6532e-04 - val_loss: 0.3805\n",
            "Epoch 73/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.6191e-04 - val_loss: 0.3810\n",
            "Epoch 74/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.5807e-04 - val_loss: 0.3822\n",
            "Epoch 75/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.5470e-04 - val_loss: 0.3827\n",
            "Epoch 76/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.5150e-04 - val_loss: 0.3837\n",
            "Epoch 77/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.4796e-04 - val_loss: 0.3845\n",
            "Epoch 78/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.4492e-04 - val_loss: 0.3851\n",
            "Epoch 79/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.4234e-04 - val_loss: 0.3859\n",
            "Epoch 80/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.3898e-04 - val_loss: 0.3865\n",
            "Epoch 81/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.3668e-04 - val_loss: 0.3879\n",
            "Epoch 82/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.3410e-04 - val_loss: 0.3881\n",
            "Epoch 83/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.3122e-04 - val_loss: 0.3887\n",
            "Epoch 84/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.2877e-04 - val_loss: 0.3896\n",
            "Epoch 85/100\n",
            "26/26 [==============================] - 0s 11ms/step - loss: 1.2635e-04 - val_loss: 0.3903\n",
            "Epoch 86/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.2436e-04 - val_loss: 0.3907\n",
            "Epoch 87/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.2202e-04 - val_loss: 0.3917\n",
            "Epoch 88/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.1962e-04 - val_loss: 0.3921\n",
            "Epoch 89/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.1792e-04 - val_loss: 0.3930\n",
            "Epoch 90/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.1577e-04 - val_loss: 0.3937\n",
            "Epoch 91/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.1376e-04 - val_loss: 0.3943\n",
            "Epoch 92/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.1189e-04 - val_loss: 0.3946\n",
            "Epoch 93/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.1014e-04 - val_loss: 0.3952\n",
            "Epoch 94/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.0845e-04 - val_loss: 0.3958\n",
            "Epoch 95/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.0671e-04 - val_loss: 0.3962\n",
            "Epoch 96/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.0491e-04 - val_loss: 0.3969\n",
            "Epoch 97/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.0334e-04 - val_loss: 0.3973\n",
            "Epoch 98/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 1.0165e-04 - val_loss: 0.3979\n",
            "Epoch 99/100\n",
            "26/26 [==============================] - 0s 13ms/step - loss: 1.0016e-04 - val_loss: 0.3984\n",
            "Epoch 100/100\n",
            "26/26 [==============================] - 0s 12ms/step - loss: 9.8455e-05 - val_loss: 0.3988\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "import cv2\n",
        "import os                                                      ## package for file path system\n",
        "import numpy as np\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras.utils import to_categorical            ## encode numbers to one-hot encode form\n",
        "from tensorflow.keras.optimizers import SGD\n",
        "from sklearn.preprocessing import LabelEncoder                 ## encode string labels to numbers\n",
        "\n",
        "path_img = 'Sign-Language-Digits-Dataset/Dataset/'\n",
        "\n",
        "images = []                                                   # input image list\n",
        "labels = []                                                   # image label list\n",
        "\n",
        "def load_images_from_folder(folder):             # function to read image files from a given folder\n",
        "    for filename in os.listdir(folder):          # for each image file in the folder\n",
        "        img = cv2.imread(os.path.join(folder, filename))  # read an image file\n",
        "        img = cv2.resize(img, (32, 32))\n",
        "        images.append(img)                         # append the image data\n",
        "        label = folder.split(os.path.sep)[-1]      # get a label for the image\n",
        "        labels.append(label)                       # append the label\n",
        "\n",
        "def convert_labels(labels):\n",
        "    enc = LabelEncoder()\n",
        "    P = enc.fit_transform(labels)                ## convert string labels to numbers\n",
        "    Y = to_categorical(P)                        ## convert number to one-hot-encode form\n",
        "    return Y\n",
        "\n",
        "for i in range(10):\n",
        "    folder = path_img + str(i)\n",
        "    print('Reading images from path: ' + folder)\n",
        "    load_images_from_folder(folder)\n",
        "\n",
        "X = np.array(images)/255.0\n",
        "Y = convert_labels(labels)\n",
        "\n",
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = 0.2)\n",
        "print('training data : ' , X_train.shape)\n",
        "print('test data : ', X_test.shape)\n",
        "\n",
        "###############################################################################################\n",
        "from tensorflow.keras.layers import Dense, Flatten\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.optimizers import SGD\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "cnn = Sequential()\n",
        "cnn.add(Conv2D(32, (3, 3), activation='relu', input_shape=(32, 32, 3)))\n",
        "cnn.add(Conv2D(32, (3, 3), activation='relu'))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "cnn.add(Conv2D(64, (3, 3), activation='relu'))\n",
        "cnn.add(Conv2D(64, (3, 3), activation='relu'))\n",
        "cnn.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "cnn.add(Flatten())\n",
        "cnn.add(Dense(50, activation='relu'))\n",
        "cnn.add(Dense(10, activation='softmax'))\n",
        "\n",
        "ops = SGD(learning_rate=0.1)\n",
        "cnn.compile(loss='categorical_crossentropy', optimizer=ops)\n",
        "print(cnn.summary())\n",
        "\n",
        "history = cnn.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs = 100, batch_size=64)\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "modiyUFUKW71"
      },
      "source": [
        "ทำ transfer learning จากโมเดล ResNet50 และ VGG19 พร้อมแสดงกราฟเปรียบเทียบ loss ระหว่าง 2 โมเดล"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JwKzNMArK0H2"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.applications import VGG19, ResNet50\n",
        "from tensorflow import keras\n",
        "\n",
        "inputs = keras.Input(shape=(32, 32, 3))             ### specify input shape for pretrained CNNs\n",
        "\n",
        "######################################   load pretrained VGG19 ##################################################\n",
        "base_VGG19 = VGG19(weights='imagenet',  # Load weights pre-trained on ImageNet.\n",
        "                   input_shape=(32, 32, 3),\n",
        "                    include_top=False)  # Do not include the ImageNet classifier at the top.\n",
        "base_VGG19.trainable = False             ## freeze pretrained VGG layers\n",
        "\n",
        "x = base_VGG19(inputs, training=False)\n",
        "x = keras.layers.GlobalAveragePooling2D()(x)\n",
        "x = keras.layers.Dense(100, activation = 'relu')(x)\n",
        "outputs = keras.layers.Dense(10, activation='softmax')(x)\n",
        "\n",
        "model_VGG = keras.Model(inputs, outputs)\n",
        "model_VGG.compile(loss='categorical_crossentropy', optimizer=\"sgd\")\n",
        "\n",
        "######################################   load pretrained VGG19 ##################################################\n",
        "\n",
        "base_ResNet50 = ResNet50(weights='imagenet',  # Load weights pre-trained on ImageNet.\n",
        "                   input_shape=(32, 32, 3),\n",
        "                    include_top=False)  # Do not include the ImageNet classifier at the top.\n",
        "base_ResNet50.trainable = False\n",
        "x = base_ResNet50(inputs, training=False)\n",
        "x = keras.layers.GlobalAveragePooling2D()(x)\n",
        "x = keras.layers.Dense(100, activation = 'relu')(x)\n",
        "outputs = keras.layers.Dense(10, activation='softmax')(x)\n",
        "\n",
        "model_ResNet = keras.Model(inputs, outputs)\n",
        "model_ResNet.compile(loss='categorical_crossentropy', optimizer=\"sgd\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "sxCYnyrJd3V3",
        "outputId": "29a1bf94-cf1b-464f-8b65-e2e27fe08144"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "history = model_VGG.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs = 100, batch_size=64, verbose = 0)\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 280
        },
        "id": "ERwo7aBlfOFl",
        "outputId": "40804c6f-8105-4071-f858-de1a6e144191"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEHCAYAAACjh0HiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3xV9fnA8c9zR3ZIIAkBEiCATEFWRBAEFQc4qlbFiasVbf21atWqraNbrUrVOqgDR6UuxIWKgIqCoAgRBMIMIIQZVkhCdp7fH+cGkpCwTwI5z/v1yiv3nvO99zwnF85zv/OIqmKMMca7fA0dgDHGmIZlicAYYzzOEoExxnicJQJjjPE4SwTGGONxlgiMMcbjAm69sYi0Bl4DkgEFnlfVJ2uUiQNeB9qEYnlMVV/e1/smJiZqWlqaKzEbY0xjNXfu3C2qmlTbPtcSAVAG3KGqGSISC8wVkSmqmlmlzC1ApqqeLyJJwFIRGaeqJXW9aVpaGnPmzHExbGOMaXxE5Ke69rnWNKSqG1Q1I/Q4D1gMpNQsBsSKiAAxwDacBGKMMaae1EsfgYikAb2B72rsehroCqwHFgC3qmpFLa8fJSJzRGROTk6Oy9EaY4y3uJ4IRCQGeBe4TVV31th9NjAPaAX0Ap4WkSY130NVn1fVdFVNT0qqtYnLGGPMIXKzjwARCeIkgXGqOqGWItcDD6uz4NEKEVkFdAFmH8xxSktLyc7Opqio6LBjPtpFRESQmppKMBhs6FCMMY2Em6OGBHgJWKyqo+sotgYYCkwXkWSgM7DyYI+VnZ1NbGwsaWlpOIdtnFSVrVu3kp2dTbt27Ro6HGNMI+FmjWAgMBJYICLzQtv+gDNUFFUdA/wVeEVEFgAC3K2qWw72QEVFRY0+CQCICAkJCVg/iTHmSHItEajqDJyL+77KrAfOOhLHa+xJoJJXztMYU388M7O4qLScjblFlJXvNSjJGGM8zTOJoLi0nM15RZRVHPkb8ezYsYNnn332oF93zjnnsGPHjiMejzHGHAzPJAJCTSpu3JGtrkRQVrbvuXGffPIJ8fHxRzweY4w5GK4OHz2aVDatu3FnznvuuYesrCx69epFMBgkIiKCpk2bsmTJEpYtW8aFF17I2rVrKSoq4tZbb2XUqFHAnuUy8vPzGT58OIMGDWLmzJmkpKTwwQcfEBkZeeSDNcaYGhpdIvjzR4vIXF9z3hqUVyhFpeVEhvnxHWSHa7dWTXjw/OPr3P/www+zcOFC5s2bx7Rp0zj33HNZuHDh7iGeY8eOpVmzZhQWFnLiiSdy8cUXk5CQUO09li9fzhtvvMELL7zAiBEjePfdd7n66qsPKk5jjDkUjS4R1KlqjcDlgTf9+vWrNs7/qaee4r333gNg7dq1LF++fK9E0K5dO3r16gVA3759Wb16tbtBGmNMSKNLBHV9cy8oLiMrJ592idHERrg7Kzc6Onr342nTpjF16lRmzZpFVFQUp556aq0zoMPDw3c/9vv9FBYWuhqjMcZU8kxnsZt9BLGxseTl5dW6Lzc3l6ZNmxIVFcWSJUv49ttvj3wAxhhzGBpdjaAugnujhhISEhg4cCDdu3cnMjKS5OTk3fuGDRvGmDFj6Nq1K507d6Z///5H/PjGGHM4xI0Lo5vS09O15o1pFi9eTNeuXff5uqLScpZtyqNNsyjio8LcDNF1B3K+xhhTlYjMVdX02vZZ05AxxnicdxJBZdMQlgmMMaYq7yQCqxEYY0ytvJMIQr8tERhjTHXeSQRiTUPGGFMbDyUC57fVCIwxpjrvJILQbzfywKEuQw3wxBNPsGvXriMckTHGHDjvJAIRBKnXZagPhCUCY0xD88zMYnCah9yoEVRdhvrMM8+kefPmvP322xQXF3PRRRfx5z//mYKCAkaMGEF2djbl5eXcf//9bNq0ifXr13PaaaeRmJjIl19+6UJ0xhizb40vEXx6D2xcUOuudiVlBHwCAf/BvWeLHjD84Tp3V12GevLkyYwfP57Zs2ejqvzsZz/j66+/Jicnh1atWvHxxx8DzhpEcXFxjB49mi+//JLExMSDi8kYY44QzzQN1ZfJkyczefJkevfuTZ8+fViyZAnLly+nR48eTJkyhbvvvpvp06cTFxfX0KEaYwzQGGsE+/jmvmbDTmIjAqQ2jXLt8KrKvffey0033bTXvoyMDD755BPuu+8+hg4dygMPPOBaHMYYc6A8VSMQ3F+G+uyzz2bs2LHk5+cDsG7dOjZv3sz69euJiori6quv5q677iIjI2Ov1xpjTENwrUYgIq2B14BknD7a51X1yRpl7gKuqhJLVyBJVbe5E5M7iaDqMtTDhw/nyiuvZMCAAQDExMTw+uuvs2LFCu666y58Ph/BYJDnnnsOgFGjRjFs2DBatWplncXGmAbh2jLUItISaKmqGSISC8wFLlTVzDrKnw/crqqn7+t9D3UZaoBlG/MID/pomxC937JHM1uG2hhzsBpkGWpV3aCqGaHHecBiIGUfL7kCeMOteABwqUZgjDHHsnrpIxCRNKA38F0d+6OAYcC7dewfJSJzRGROTk7OIcfhc2kegTHGHMtcTwQiEoNzgb9NVXfWUex84Ju6+gZU9XlVTVfV9KSkpFrf4MCauNyZWVyfjvX4jTFHH1cTgYgEcZLAOFWdsI+il3MYzUIRERFs3bp1vxdJt2YW1xdVZevWrURERDR0KMaYRsTNUUMCvAQsVtXR+ygXBwwBrj7UY6WmppKdnc3+mo225BdToVCyJfxQD9XgIiIiSE1NbegwjDGNiJsTygYCI4EFIjIvtO0PQBsAVR0T2nYRMFlVCw71QMFgkHbt2u233A2vfE9OXjEf/abXoR7KGGMaHdcSgarOYM/qz/sq9wrwiltxVBXwCaXlFfVxKGOMOWZ4amZxMOCzRGCMMTV4KxH4hNLyY7m72BhjjjxvJQK/jzKrERhjTDWeSgQBv48SqxEYY0w1nkoEYX6hrMJqBMYYU5WnEkHA76O0zBKBMcZU5alEEPT7KK2wpiFjjKnKY4nA5hEYY0xNHksEPlSh3GoFxhizm6cSQcDvTHS2WoExxuzhqUQQ5ndO1xKBMcbs4alEEPBV1gisacgYYyp5KhEEA1YjMMaYmryVCHyWCIwxpiZvJYKANQ0ZY0xNnkoEgVCNwBaeM8aYPTyVCIKhUUMllgiMMWY3jyUCp2mozJqGjDFmN48lAussNsaYmjyVCPbMLLYagTHGVPJUIrCZxcYYszdPJYLKpiG7OY0xxuzhqURQ2TRUUmZNQ8YYU8m1RCAirUXkSxHJFJFFInJrHeVOFZF5oTJfuRUP7GkashqBMcbsEXDxvcuAO1Q1Q0RigbkiMkVVMysLiEg88CwwTFXXiEhzF+MhYH0ExhizF9dqBKq6QVUzQo/zgMVASo1iVwITVHVNqNxmt+KBPfMIbNSQMcbsUS99BCKSBvQGvquxqxPQVESmichcEbmmjtePEpE5IjInJyfnkOOweQTGGLM31xOBiMQA7wK3qerOGrsDQF/gXOBs4H4R6VTzPVT1eVVNV9X0pKSkQ45l96ghqxEYY8xubvYRICJBnCQwTlUn1FIkG9iqqgVAgYh8DfQElrkRj92q0hhj9ubmqCEBXgIWq+roOop9AAwSkYCIRAEn4fQluGLPhDKrERhjTCU3awQDgZHAAhGZF9r2B6ANgKqOUdXFIjIJ+BGoAF5U1YVuBbTnVpVWIzDGmEquJQJVnQHIAZR7FHjUrTiq8vsEEbsfgTHGVOWpmcUiQtDno8SahowxZjdPJQJw5hJYjcAYY/bwXCII+H3WR2CMMVV4LhEE/T5KK6xpyBhjKnkwEQilZVYjMMaYSh5MBD7KrEZgjDG7eS4RBPxCifURGGPMbp5LBGF+n40aMsaYKjyXCAJ+sSUmjDGmCs8lgqANHzXGmGq8lwh8lgiMMaYq7yWCgNj9CIwxpgrPJYKA1QiMMaYazyUCp4/AagTGGFPJg4lArEZgjDFVeDARWNOQMcZU5blEYPMIjDGmOs8lgjCrERhjTDWeSwQBv9iic8YYU4XnEkHQ77NlqI0xpgrPJYIwv4/SCksExhhTyXOJwDqLjTGmOs8lgqDfR3mFUmH9BMYYA7iYCESktYh8KSKZIrJIRG6tpcypIpIrIvNCPw+4FU+loN85ZWseMsYYR8DF9y4D7lDVDBGJBeaKyBRVzaxRbrqqnudiHNUE/eIEV66Eu3n2xhhzjHCtRqCqG1Q1I/Q4D1gMpLh1vAMV8IVqBDaXwBhjgHrqIxCRNKA38F0tuweIyHwR+VREjq/j9aNEZI6IzMnJyTmsWIKBykRgfQTGGAP1kAhEJAZ4F7hNVXfW2J0BtFXVnsC/gfdrew9VfV5V01U1PSkp6bDiCfqcpiGrERhjjMPVRCAiQZwkME5VJ9Tcr6o7VTU/9PgTICgiiW7GVNlZbDenMcYYh5ujhgR4CVisqqPrKNMiVA4R6ReKZ6tbMYEzjwCgxGoExhgDHGAiEJFbRaSJOF4SkQwROWs/LxsIjAROrzI89BwRuVlEbg6VuQRYKCLzgaeAy1XV1a/qYZU1Ahs+aowxwIEPH71BVZ8UkbOBpjgX+P8Ck+t6garOAGRfb6qqTwNPH2AMR0Sgch5BmTUNGWMMHHjTUOUF/Rzgv6q6iP1c5I9WlfMIbEKZMcY4DjQRzBWRyTiJ4LPQBLFj8kq6e2axrUBqjDHAgTcN/QLoBaxU1V0i0gy43r2w3LN71JCtNWSMMcCB1wgGAEtVdYeIXA3cB+S6F5Z7bNSQMcZUd6CJ4Dlgl4j0BO4AsoDXXIvKRWE2j8AYY6o50ERQFhrWeQHwtKo+A8S6F5Z7KmsENrPYGGMcB9pHkCci9+IMGz1FRHxA0L2w3LO7s9gSgTHGAAdeI7gMKMaZT7ARSAUedS0qFwV9tuicMcZUdUCJIHTxHwfEich5QJGqHpN9BMFA5f0IrEZgjDFw4EtMjABmA5cCI4DvROQSNwNzi92PwBhjqjvQPoI/Aieq6mYAEUkCpgLj3QrMLWF+axoyxpiqDrSPwFeZBEK2HsRrjyo2asgYY6o70BrBJBH5DHgj9Pwy4BN3QnKXzSw2xpjqDigRqOpdInIxztLSAM+r6nvuheWeykXnSmytIWOMAQ68RoCqvotzt7FjmogQ8Indj8AYY0L2mQhEJA+orQ1FAFXVJq5E5bKAX6yz2BhjQvaZCFT1mFxGYn+Cfp91FhtjTMgxOfLncFkiMMaYPTyaCMRuVWmMMSGeTAQBn89uVWmMMSGeTARhAZ91FhtjTIgnE0HQL7bonDHGhHgyEQR81llsjDGVXEsEItJaRL4UkUwRWSQit+6j7IkiUlZfK5oGrWnIGGN2O+CZxYegDLhDVTNEJBaYKyJTVDWzaiER8QOPAJNdjKWaoE+sRmCMMSGu1QhUdYOqZoQe5wGLgZRaiv4GZ+mKzbXsc0XQ77Ob1xtjTEi99BGISBrQG/iuxvYU4CLguf28fpSIzBGROTk5OYcdT8AvlFiNwBhjgHpIBCISg/ON/zZV3Vlj9xPA3aq6z6uyqj6vqumqmp6UlHTYMYX5fbbonDHGhLjZR4CIBHGSwDhVnVBLkXTgTREBSATOEZEyVX3fzbgCNrPYGGN2cy0RiHN1fwlYrKqjayujqu2qlH8FmOh2EoDQWkNWIzDGGMDdGsFAYCSwQETmhbb9AWgDoKpjXDz2Ptmic8YYs4driUBVZ+Dct+BAy1/nViw1OTOLrWnIGGPAqzOLrUZgjDG7eTIRhPltZrExxlTyZCII2MxiY4zZzZOJIBiwmcXGGFPJm4nA58wsVrVkYIwx3kwEfue0yyssERhjjCcTQSCUCKzD2BhjPJoIgn5neoPNLjbGGM8mglCNoMwSgTHGeDIRhPmUIb75lJWXN3QoxhjT4DyZCDplj+fVsEeQldMaOhRjjGlw3ksEZcV0W/ECAMENcxs4GGOMaXjeSwQZrxFZtIldGk5w07z9lzfGmEbOW4mgrBhm/IvtCX34tKIf4Zvng00qM8Z4nLcSwQ//hZ3rWN3jN8yvaE+wMAd2rmvoqIwxpkF5JxGUFcP0f0Hr/uS3GsSPFR2c7esyGjYuY4xpYN5JBD++DTuz4dS76dIyjhW+NMrxw3pLBMYYb3P15vVHlRNGQHgMtD+NJBHO7d2OJQva0GnNXIINHZsxxjQg79QIAuFw/EUgzvISvzylHfPK21OxLgNsqQljjId5JxHU0DE5lpLkXoSX51O8eXlDh2OMMQ3Gs4kAoM+A0wGYO+vzBo7EmKPIhvkw/hdQVtLQkZh64ulEcELvkygmnA2ZM/lg3jrufGc+pz8+jSmZmxo6NGMazvIpsHA8ZH/f0JGYeuLpRCD+IPnNjqdt8VJufXMeXyxaT9/Cb7n/7Vms21HY0OEZ0zAKcpzfq75u2DhMvXEtEYhIaxH5UkQyRWSRiNxaS5kLRORHEZknInNEZJBb8dSlWceT6BVcw5Sf+5mb/FceLXuIJ/RR7nzje8rsBveN284NNrO8Nvmbnd+rpzdsHKbeuDl8tAy4Q1UzRCQWmCsiU1Q1s0qZz4EPVVVF5ATgbaCLizHtRVL6EvjuOTp+chk0SYUB/0f/WU/z07rHefqLZP7v9I5krNnB9OU5RIcH6N4qjl4lc4jYvoyi4lKKSkrILyph565idhaW4ENpGhNBQkw4cfEJhCd3hsTjILYViFChUFRSTKC0gGBZPlK6CxBnNFNpIWxeDJsWQe5aSOoMKemQfDzkb4Ity2HHTxAWDTEtIDoRSvIhb5PzLS6pM3Q8EyLi9n3SxXkw89+wayv0uhJa9dk9muqwbVsJTdsdufdzy7ZV8HQ6XPyiM5rM7FFZI1g7G0p2QVhUw8ZjXOdaIlDVDcCG0OM8EVkMpACZVcrkV3lJNFD/X8/anQIJx0GX82DI752LbFg0l331CI999Rjpsy5lx65SfAJBLeHBwKsMCnwJQEzoJ/EgDucD9vffqsQXwc6w5jRb+ik+Pch7JviCzjl1GAptBkDLE8AfmilRUQ7z/gef/wUKNkMgEr5/EZJ7QNfzITYZohIhMt55H38AopMgvs3+j6sKUx+Eb56E0++HwXceXNz1LetzqChzLnaWCKoryHG+TBTlwtrvoMNpDR2RcVm9TCgTkTSgN/BdLfsuAh4CmgPn1vH6UcAogDZtDuCidDBiW8BvaixHfeq9lOZkcWfm25wZvYlg94G07dKLiGl/x795ET+0uZ45ra8lJjKCJlHhJMZGkpoQQ3KTKMpVWb2lgOUbc9m6ZSNsySJ850oiS7YSFfQRGfQRDAYp8kVTKFHkVYSxNb+IrfnFbCmsYIWmsqaiOSUFgpbuorusopNvHRu1KSu1JZv8LQiWF5LIDhLYST6RbNZ4dhBDT8liRMx8hq6dQ1LWFwCU+iIoCYsjWFFMoLwIX3kRpPaDK96ExI6w4B3IeBWm/aPuv1Hnc2HwHZDSt/b95WUw8TZnLafYVjDtYeg83KnJHK0q2783LmjYOI5G+Zuh0zBY+K7TPGSJoNETdbmNVERigK+Av6vqhH2UGww8oKpn7Ov90tPTdc6cOUc4ylqUFcOke50RFLlrnG2RTeHnLzjNLy5TVXYWlbF+RyEbcgvZkl/C1vwSduwqISzgIyosQHS4n5jwANHhASKCfpZu3Ml3K7cxe9U2IopzSPctI923jFh2UUQYxQT5oeI45sUMoVtKHG0ToomNCBAbESQuUE5T8miiO4isyEfLS9GyUhJyF9Jq6atIUS6knuh8U1QFn9+pLcQkw6aFsHwyDLkb+o2CZ06CuBT45edObaSiAn76BhI6QJNWrv/t9quiAh7tAIXbnM/096uO/qas+lJeCn9NhCH3QNYXgMIvpzZ0VOYIEJG5qppe2z5XawQiEgTeBcbtKwkAqOrXItJeRBJVdYubcR2QQDicN9p5nLvOGVud0sepQdQDESEuMkhcZJCuLZsc0GuGdEpi1OAOqCplFUqFKhUVkFtYyua8IjbtLEa35BO2fieZG3YyK2srBSW1NT35gPDQT3/aRqdzX6tZnLhrFmV5Gykuq0DLy0iU+YQXbUEAhj0M/X/lvPy80fD2NTDjCWg7AD77I2yYB/5wSL8eBv3OaYaqKXed83ePPpjGtkOweZGTBFL6wrq5zgq0canuHvNYURD6rxeTBO0Gw4x/OX1K4bENG5dxlWuJQEQEeAlYrKqj6yhzHJAV6izug3Pl2epWTIcsLsX5OUaICEH/nm+4kWF+WsRFhJ5VvwCXlVeQX1xGXlEZu0rKKSgpo7i0gqBf8PuEDblFfDR/PbcsGUJJ2SkAhAd8RAT95BaW0jw6wCW9mtM1Mpl263Jp3SwKTRtOWKcLifzy7wjqdMKf/6QzLn32CzD3VRh0m5MQAmFOIJkfwns3O/0S5/0Lul+8J8gty53+jeZHaBzByq+c3wNugfE3wMaFlggqVXYURzd3+s6mPwY/zYJOZzVsXMZVbtYIBgIjgQUiUnkrsD8AbQBUdQxwMXCNiJQChcBl6nZblakm4PcRHxVGfFRYrft7A+f0aEleUSlLN+aR0jSS5NgIKlSZtjSHN79fy39mbaD8m/XVXhfPcJ4KZpEV3Yu0M+/i1O5toM+1ZHb4Jf4v/0aXaQ+hmR8gFzwNSz5xLjgpfQFxLs5LP3U6vDNegzUzndrEZa8fmQvSqq+di1zH0HttWgCdhx3++zYGBaGhozHNoWVP8IfB6q8tETRybo4amgHss+FVVR8BHnErBnPkxEYESU9rtvu5D+GMbsmc0S2ZwpJyVm8tYNWWArK37yLg8xEd7md1WX9embGKn8YtpGfqWrbtKmHttkLgBs7w9eTRrS/T9AVnmQ96Xw3njgbxw/TH4atHnI7sZu1h6IOQ+QG8eSVc+rIzwulQlZc6/RUnjHCaO5qmOTWCo40qLPnY6Y8KhNffcfMrawRJEIx0BhbYxLJGzzvLUBvXRIb56dqySa19GVf0a8M7c7IZ+80q2iXG8NvTOzK0azITMroyfMrx3MR44tp0p2mnXzJAA0QE/HDq3XD8hbBrG7Q+CXw+SL8Bxl0Cb18LF41xLuSHYv08Z+5Fu8HO8xY9js6RQ5nvwzvXwYVjoNcV9XfcyhpBdJLzu91gmPYQFG53OtZNo2SJwLgq6Pdx5UltuPKk6sN+f3lKe847oRX//KwDkxZuZNeKOUQG/bRPiiYs4CM84KN7qzh+nVhGs+gwZ27DyPfgf5fBhBthyUQY9gg0abn3QcuKnZFMtU2UWxXqH0hz+jtI7gGLJ0JJgTOH5Gig6tSKwOnYrk/5myEQsadzuFUvQGHLCmh9Yv3GYuqNp9caMg2rRVwEo0f0IuP+M3n1hn6MSE+lRZMIYsIDlJUrL89czZB/fslz07IoKi13Lk4j34PT7oOlk+DpE+Hb55wLf6XcbBg7DF44Habcv/cSEqu+ci7+lSOTWnQHFDZlctRYMdWppYjPmWlenwq2OB3FlQm0SWiQxM7s+o3D1CurEZgGFxH0M6RTEkM6JVXbvmJzHg9/uoRHJi1hzFdZDO6UxGmdk2iRej3Le/elz4K/0mPSPRR+9SQRp9+FxLeF90Y5yyd3Oc9ZRqOkAM553Gle2rUN1nwHJ/5yz0GSuzu/Ny3Y84137fdO30FM9XjqzfTHIa6103meXQ9zZqoq2Fz9vCtHU+VaImjMLBGYo9ZxzWN58doT+XblVsbPzWba0hw+mu+MThKBLsl/pofM5fKC1+nz8e8A0KQuyGWvO6OCpv4JvnnCuYiVFsJPM0HLnVnPleLbQHjcnn6CtbPhpbOc+SJXvAGtetfvSf80E9bMguGPQvFOp6+gaCdEHNhcksOWn1N9KG1EHITFWiJo5CwRmKNe//YJ9G+fQEWFsmj9TrbkF9OnTVPiooKUlQ/inTk/55XJb9OuaDFfl17ORcuDDA8vIbfnnfh2+khb+CSa0BH/wFuh63nVl8oQcZqHNi50ahIf/hZiWzozp8cOdzqmj7+w/k52+uNOR22fkZDlrGlFzhJo3a9+jl+wOdQvECLizKGxRNCoWSIwxwyfT+iRWn1l1YDfxxUnteWC3rczIWMdpd+v4YEPFvHAB5WdrOkEeZmwzeFc2a4NN8S0Y6/u5eTu8MPrMGM05CyGK95yZpG/eRW8cy0UPAb9bnT/BDdlOv0DQx90hm5WTqDbvLh+EkFFhdNHENO8+va4VEsEjZwlAtMoRIUFuLp/W67u35YF2bnMzNpCcpMIWjeLQgRenbmasd+s5pWZq7n9zE7cPLgDPl+oQ7RFdygtcOYudL94z+Sy6yY6w1Un3eNMrnL7YrxkIiDQe6TzPD7NWSE2Z4m7x61UuN1pOouukQiapDhLrJhGyxKBaXR6pMbtVXPo06Ypd57VmYc/XcI/Jy1lVtZWRo/oRUTQx7y8lpwCFAdiKR/69z3LhAfCnaah54c4Y/pvmg7RCe4FvuwzpyZS2Vnr8zn3mNhcTyOads8qrtFJHtfaWXqitAiCEXu/zhzzbPio8YzWzaJ4+srePPTzHsxetY3TH59Gn79O4RefFjBfO3B7wbWc9MSP3P/+QrK373JeFBkPl77qXAjfu8lpPqmpZJezdv/hKNjiLIDX8ezq25t3hc31VCPIrzGZDKioULI1NJFs57r6icPUO0sExlNEhCv6teHD/xvEKR0TuWFgO16/aTDHPzCHG0bdzpndknlrzlrOeXI6kxZucF7Uqhec/Q9YMQU+vctZjbPSymnw7z7O0ts71h56YCumArr3EufNu0L+Rmfoq9uqLjgX8tGP67lrcujY1k/QaFkiMJ7UuUUsz17Vl3vP6Uq/ds0I+H2kpzVj9GW9mHr7ENISo7n59Qwe/GAhBcVlztyDfjc5d3R7qg/MfQUm3w+vXUixL4ryonwYdykU7ji0gJZ95lyAW/aqvj2pq/O7PvoJ8qssOBeS8dN21muoOcxqBIenOB/mjD0q75NticCYGtokRDH+5pP5xaB2vDrrJ7r/6TOGPDaNUVtG8L8TXmFbRCp8dCvMfIpJkcPpuek+bii6lYqtK+Ctq6vPdK6irLyWZiVw7vCW9bmzGqqvxn/JqiOH3FaQA74ARMTv3rRgXXAu1igAABbxSURBVC4bNbTYoNUIDk/m+zDx9qOy4906i42pRVjAx/3ndWN49xZ8s2IrSzftZMmGPKYuDuMPegdn+uZSRBg/RfXnd+e04b/fxvNA4a/42+qn4IWhENXUWek0EE5OZAfeyY5jUm4bHv/VJXRMrnGTl7XfOX0MtS31HNcawmIOq0bw1vdrKCgu54ZB7fZdsGCz0z8QSkZl5RVkbthJMWHk+uKJs0RweLb/5PzOXVt9rsZRwBKBMfuQntas2vLbRaXlLN+Uz5KNPUmMCWdwpyT8PmFo12Qufq6MhLASfiPfUlFUSEGFn52bN9K8cCa/lhJuRvjwxc9ofsuzxMXveU+Wf+Z8E29/6t4BiEBSl0OuEZSUVfDQp0soLCnnkvRUmkQE6y6cn1Pt7nBZOQUUlVYQHeZnnSZYIjhcuaE+pNyjr4nNEoExByEi6K91eGqHpBheuCadq14s55m8IZRVOO3AkUE/owa14cbukD/jeX6W+Qo7/n0i5Rc9hr/LOc4Q1WWToc0AZzmH2jTv4iyydwimL89hx65SACbO37DXKrDVFGyu1lG8YJ0zEur8nq1YM68ZnXZk2wXjcFQOJsg9jEEFLrHP1Zgj5MS0Zoy99kSmZG6kQ/MYOiXH0rVlE+IinW/hMZc9waTPzqDtN/fQbPx1VAQi0NR++HMWs6Tl3Yx58wcy1uzgzG7J3DCoHSnxkc4bN+/mzHzOz9lrjH9+cRlhfh9hgdq7+96ft574qCCJMeG8M3ftvhNBfo5T+whZuC6XqDA/w3u0JOuHBCQ30+norLm0tzkwO9Y4v4/CmpUlAmOOoEEdExnUMbHO/cPOPo8/5qewae5HDCxbyMkrF5Eq4dz4fQvyInPokRLHqzNX8+rM1Zx3QkvOPaEVA+M7OpPcchZDTBJl5RVMW5rD23PW8sWSzfh9Qs/UePq0bcoV/VrTNsG5r0J+cRlTMjdycZ9U0hKi+fsni1mxOY/jmtdyI3pVp7O4yhyCBetyOb5VE45v1YTpmoC/rMDpy4iM3/v1Zt/Ky/aMurJEYIz560W9mN27DdnbC5m8o5D8olL+3jGJAR0SCPp9rNtRyNgZq3hz9hren7eeZF8uM8N8ZP7vPu4Ou4fVeT52lZSTGBPOdSenATBn9TbWfzOOP89vy5O3XUtsRJApmRspKq3gwt4ptE2I4uFJS3hnbjb3Du+6d1DFO6G8eHciKK9QMtfv5LITW5MYE05eeAuowLmIWSI4eHkbnOU7fEFLBMYYZ/G8/u3rXqoiJT6S+8/rxu+HdSbjpx3MWJHDs5m/49e5oxnje5C3eo6mV9dOnNo5iaDf54xP/+gx2PIuWwqbMHpCOx68cijv/7CelPhI+rZpis8nnNY5iQkZ67jrrM4E/DWakirvVRyaQ5CVk09haTk9Upx+i8jENrAZ5yLWorsbf5bGrbJfoFVvyJ7tDDGuz3tR74fNIzDmKBUe8DOgQwJ3nd2F39x+P/4r36RN+VruWvdbziz6jODKL5wby79wOix6D/rfQpy/mDMW38/rs1YyY8UWftar1e7F9S7p25qcvGK+Xp6z98Fq3Kt4QbbTUVzZKd6sVXsAyg9n9rSXVfYPtD3Z+b1zfcPFUgtLBMYcKzqdDdd84LTTf/gbGHcxvHo+7Nrq3MJz2D+Qcx5loH8RGyY+RHmFckGvVs6SGCW7OL1Lc5pFh/HKzJ8orTG5rSKv+qziBetyiQz66ZAUA0Bq6zRK1E/uxlX1esqNRmUCbTPA+X2UzdK2piFjjiVtToLfLXHanPM2Ot/kU/tBbDIAgb7XkL/0C3637B06RhfSZeK/YF0GhMcQNvgubhp4Bg9NXsll/5nFU5d2I9W3jR9Xb2DplI+4FFiYG073Fs6IoW6tmuAP1Sa6tIxnkzZDtqyh2T7Cqw/fr97GhIxs/nZhj93xHfVy1zi1rcSOoedHVz+Ba4lARFoDrwHJgALPq+qTNcpcBdwNCJAH/EpVj77518YcTQJh0LSt81OTCDEX/5uiZ+dzQd4nQF8YFFrWYPJ93BT/Amd1G0hu1vc0f2YVUMYJwAlAIeFc+2YWr49qwaJQR3Gl45rHMI8EWjZw09COXSX83/8y2LSzmOHdWzK4UwPdV/pg7Vjr3Ba1SSvn+VE2l8DNGkEZcIeqZohILDBXRKaoatXF1VcBQ1R1u4gMB54HTnIxJmMav4gmRNwyA7Si+r2Os76AKQ/Qbs17FKWcwMTtJ/B9QTKnHN+WoSeksT3YisDbmxkxZhaFpeV0T9kzwS0s4CM/vAURhVWWuti1DaKOUP3gAOYnqCp/fG8h2wpKiAkPMH5u9rGTCHLXOnfCC0ZCVKJ3agSqugHYEHqcJyKLgRQgs0qZmVVe8i2QijHm8IXH7L2tw+nOT0UFET4fF1YoZ5eWEx3uXAZaAa//Io8R/5kFsHvEUKXyJik03Toddm1jx4f3EL/kLdZ2vp7Enz9KZHiQ3F2lfDB/HZ8s2MA1A9I4p3sLWPmlcwGMac6khRtJiAnjxCpLduSsX03Ba5dTEplEp99+uM9k8N4P6/h4wQZ+P6wzG3OLeOv7teQWlu6esHegVBWpz0lxFRVOjaDzcOf5/m79WbjdWfivHmOslz4CEUkDegPf7aPYL4BP63j9KGAUQJs2+5gZaYzZv9Cicj6f7E4ClTomx/LfX5zExB830LF59WQSntCGwNZydj7ak5iKfGZWdOPkpS/z0d+XMq7FPWSs30VJWQVNIgL85Y0vOTFtPEnrP0cTO/Fc+2f559dOh/RpnZO46+wuZC+ZzQlfj6KNbsNXtJisWe/R4eSf1xpy9vZdPPjBIvqlNeOmwR1YtD6X12b9xMc/7mfZjFre58JnvuEvF3TnnB573b3aHQU5zhyNuFCccamwdUXtZXeuhyd7OXfG617738INro8aEpEY4F3gNlXdWUeZ03ASwd217VfV51U1XVXTk5KOkaqgMceo7ilx3DO8y557Oockp3UDYLMk8FG/12l/xxes6nkn5/tmcs/2+3moQyZfn7Od2WetZnL43TRZ9zXrulxH+dbV9Pv211zRO4m7h3Vh7k/beOLp0Qz46mqCPmHlRR+yhhYEv/gTWl66VzxFpeX8elwGCjw+oid+n9AjJY6OzWMYP/fg2tr/NnExW/JLePP7emyjr+wPiK9MBK2dGkFt9yVY+ZWTNJZ9Vn/x4XKNQESCOElgnKpOqKPMCcCLwHBV3epmPMaYQ9d5wPlsjfuADp1P5rhAmLPxovshrT29PvotvX6aD6GVlgMtevOrghuZOj+eYRLHs2FP0bfiCSTxEm5MfoLApvlsb9KZpje8R0J8Cl+vvpPB8+5k4Sdj6H7+b3YfU1X5/fgfWbAulxeu6knrDZPho7HI9tWM7P4CD3yxhayc/N3DXPfl62U5TFq0kZZxEXyzYgvbCkpoFh3mxp+quso5BPGhzve4VCjJr325jtUz9vyux3WdXKsRiNMI9xKwWFVH11GmDTABGKmqy9yKxRhzBIiQcPypSKDGxbP3VXDncvhNBvz6O/jVLAI3TuUfoy7hxLRmHH/GSBj2MLL0Yxh/PYHSfDj/SZr+djq++BQATj7vBhb5u9Ai43FKdu25FehzX2Xxyfw1vNZtLmdMOh3euRa2rYS8jYzI+Tc+gQkZ++94LSmr4E8fLaJtQhTPXNWH8grls0UbD+vPMTVzE099vpxtBSX7LliZCOKqJAKovZ9g9XRnGYqd2bC9/uZsuFkjGAiMBBaIyLzQtj8AbQBUdQzwAJAAPBvqvClT1XQXYzLGuCGq2V4jiJJiA7x9U2gCFR2dZbbDY6DzOeDzVysbCPgpOu3PJE69jNkv3syatpewuCyFzLnT+LrJOFplrXbu1/Czf8NxZ8A3TxDx+V+4PbUP4+ZGcN3J7UiKrXvJhpe/WcXKnALGXpdO79bxtEuM5uMfN3BFv0Prc3xj9hr+8N4CVGHMV1mMHNCWUae0JyGmegxzf9rOjm/ncEqwCRqIJhx2J4INa5fTsupyHTvWwI6fIP0G55aWq2dAs/aHFN/BEj0K75+5L+np6TpnzpyGDsMYc4SpKjMeG8EpBZOrba+Ib4tv2MPOqJvKppLyUnjhNIp3bOSk3L9TFGjCVSe15abB7WneJGL3a4tKy3n7u1V8N/lNSBvEMzecCsBjny3l2WkrmP3HM0gMXbxXbykgKyef4tJy0jKfIzalE6mDrq4+wihvI1MmT2Ti3Cx6tgjnlMFD+ffiGD76cT1JMeF8dttgmkb6oSCHsqjmnPvUDH6/7QFayDauDRvNWccnk7P+J17IuYr7Sq+nz8V38vM+oRrCvDfg/Zvh5hnw3587ie/iF3b/bbJy8gkP+GndLOqQ/r4iMreuL9o2s9gYc1QQEfrd9gY5G1YSl7eMsK2LIRiFL/0GZ/x9Vf4g/Oxpwl84nendP+Xl0jP5YtYKrpr5Df6kTvRIjadZTBgzvv+BB8ue4BrfUsq2jYPl/4aOZ3Bez5Y8/eUKJi3cyNX92zIzawvXjf2ekvIKfuX/kHOCb1K+WLhv1mZS+11IatNItq5eyEU/juLM8u2cGQZsAyY+x1O/msn1A9O4dMws/joxk9FhY2DhBD4d8CZLNxXSr3k+xU06ckIgjvFzsunWMo5yCdC7ST5/+nARJ3dIpEVchFMDiGwKzY+HtEHo6hlMXriBqYs3M2PFFjbkFnHjKe3447ndjvzf3moExphj1tQ/wYx/VduU42/BJxUnsbQkkT+GvUVEAHxD7kLmv+Hc+7nXVeig33HGq9kkxYZz37nduPz5b2kVH8FzfdfT4YubyetwPiU5K4jKW8UlRQ9QQDhvhf2VgMAnnf/OlWf0x19eBC8Pd1YUveYDRk9ZxoJp7/By2KMAzKMLD7UYzZvbRiC9R8Lwh/fMYXiyJ/mJvUhfcjkD2icw9roTkad6OXMuLh9HwTfPEz3lLoYUj2ZHRGsGHpfAoOOSGNwpkdSmViMwxpg9hj4InYZDSZ5z85f8TSQt/ohrVn6MBMugZR+45CWnrb3/r+CrR2DGE8i8cYyPaMO7a47njRc7MiA8hYfP6kzC+7+DlHRiL3/BGdXzwlA+jHqSCoSABpDrJjKyeZX7OZzxJ/j4dzD/TW45eRg7Z41lJa1ZkHIFF2T/kydTpyEb83ePGNrdzBTXmpiijdwzrAt/+iiTj2d8z3nbV1Pe72ZmLMvh6WnhvAP8red2Th5xg+trKlkiMMYcu0Schfiq6nstsmsbbFoIrfs7azOBs/7/0Aeg7/WwbBJhCyYysnAq4foplADvAE1S4PL/OU1RwUi46h38Y8/G7/PDtROheY2b+vS9Hn58Cz77A+HLJ5PIdm4svpV5KzrQrVkfOs59zCkX17r66+JSYdV0rhmQxqcLNzJjyvuc54MLJwoLymfTPrENpYEkTgkugXpYWM8SgTGm8YlqBu0G174vvjX0u5Hofjfy9qzl9GuaR5pscoZzdjxz90quACR3g5u+An/YnmGfVfl8cP6TMOYUWDQBOfm3nFR2Fqu+X0v8iGfgv6c6E8Tia4xOikuFvA34tJzHLu3Jkv88Tn5pLCeffArXt4xjWPcWBD8cUm/zCSwRGGM8a8SAjvsvtL8hnM27wpl/gaWfwGl/4N5gJLed0YnIMD+c/keY9jA0a1f9NU1SnFtXbllK6+bdaB21HJKHcO+5x+8pkzYIFo6HrVmQeNzBn9xBsM5iY4xxU0kBhEVX37byK3jtZ87jQASUFcGwh51+jEpbVsDTfSEQCeJzVpM9+f/g9PsOKQzrLDbGmIZSMwkApJ0CV70L27KctYgKt0P3i6uXSegAZ//DabISn/OT2s+VEC0RGGNMffP5oOMZwBl1lxGBAbfUTzj1chRjjDFHLUsExhjjcZYIjDHG4ywRGGOMx1kiMMYYj7NEYIwxHmeJwBhjPM4SgTHGeNwxt8SEiOSw+xbZBy0R2HIEwzlWePG8vXjO4M3z9uI5w8Gfd1tVTaptxzGXCA6HiMzx4j2RvXjeXjxn8OZ5e/Gc4cietzUNGWOMx1kiMMYYj/NaIni+oQNoIF48by+eM3jzvL14znAEz9tTfQTGGGP25rUagTHGmBosERhjjMd5JhGIyDARWSoiK0TknoaOxw0i0lpEvhSRTBFZJCK3hrY3E5EpIrI89LtpQ8fqBhHxi8gPIjIx9LydiHwX+szfEpGwho7xSBKReBEZLyJLRGSxiAzwwmctIreH/n0vFJE3RCSiMX7WIjJWRDaLyMIq22r9fMXxVOj8fxSRPgdzLE8kAhHxA88Aw4FuwBUi0q1ho3JFGXCHqnYD+gO3hM7zHuBzVe0IfB563hjdCiyu8vwR4F+qehywHfhFg0TlnieBSaraBeiJc+6N+rMWkRTgt0C6qnYH/MDlNM7P+hVgWI1tdX2+w4GOoZ9RwHMHcyBPJAKgH7BCVVeqagnwJnBBA8d0xKnqBlXNCD3Ow7kwpOCc66uhYq8CFzZMhO4RkVTgXODF0HMBTgfGh4o0qvMWkThgMPASgKqWqOoOPPBZ49xiN1JEAkAUsIFG+Fmr6tfAthqb6/p8LwBeU8e3QLyItDzQY3klEaQAa6s8zw5ta7REJA3oDXwHJKvqhtCujUByA4XlpieA3wMVoecJwA5VLQs9b2yfeTsgB3g51Bz2oohE08g/a1VdBzwGrMFJALnAXBr3Z11VXZ/vYV3jvJIIPEVEYoB3gdtUdWfVfeqMF25UY4ZF5Dxgs6rObehY6lEA6AM8p6q9gQJqNAM10s+6Kc6333ZAKyCavZtPPOFIfr5eSQTrgNZVnqeGtjU6IhLESQLjVHVCaPOmympi6PfmhorPJQOBn4nIapxmv9Nx2s/jQ80H0Pg+82wgW1W/Cz0fj5MYGvtnfQawSlVzVLUUmIDz+Tfmz7qquj7fw7rGeSURfA90DI0sCMPpXPqwgWM64kLt4i8Bi1V1dJVdHwLXhh5fC3xQ37G5SVXvVdVUVU3D+Wy/UNWrgC+BS0LFGtV5q+pGYK2IdA5tGgpk0sg/a5wmof4iEhX691553o32s66hrs/3Q+Ca0Oih/kBulSak/VNVT/wA5wDLgCzgjw0dj0vnOAinqvgjMC/0cw5Oe/nnwHJgKtCsoWN18W9wKjAx9Lg9MBtYAbwDhDd0fEf4XHsBc0Kf9/tAUy981sCfgSXAQuC/QHhj/KyBN3D6QUpxaoC/qOvzBQRnZGQWsABnVNUBH8uWmDDGGI/zStOQMcaYOlgiMMYYj7NEYIwxHmeJwBhjPM4SgTHGeJwlAmPqkYicWrk6qjFHC0sExhjjcZYIjKmFiFwtIrNFZJ6I/Cd0r4N8EflXaC38z0UkKVS2l4h8G1oH/r0qa8QfJyJTRWS+iGSISIfQ28dUuY/AuNAMWWMajCUCY2oQka7AZcBAVe0FlANX4SxwNkdVjwe+Ah4MveQ14G5VPQFnVmfl9nHAM6raEzgZZ5YoOKvC3oZzb4z2OGvlGNNgAvsvYoznDAX6At+HvqxH4izuVQG8FSrzOjAhdF+AeFX9KrT9VeAdEYkFUlT1PQBVLQIIvd9sVc0OPZ8HpAEz3D8tY2pnicCYvQnwqqreW22jyP01yh3q+izFVR6XY/8PTQOzpiFj9vY5cImINIfd94lti/P/pXKFyyuBGaqaC2wXkVNC20cCX6lzh7hsEbkw9B7hIhJVr2dhzAGybyLG1KCqmSJyHzBZRHw4qz/egnPzl36hfZtx+hHAWQ54TOhCvxK4PrR9JPAfEflL6D0urcfTMOaA2eqjxhwgEclX1ZiGjsOYI82ahowxxuOsRmCMMR5nNQJjjPE4SwTGGONxlgiMMcbjLBEYY4zHWSIwxhiP+38HR+QMQaa3vQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "history = model_ResNet.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs = 100, batch_size=64, verbose=0)\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ys8sxEwngP_y",
        "outputId": "817e06b7-14a8-4008-859f-4406db214416"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "13/13 [==============================] - 0s 5ms/step - loss: 0.4328\n",
            "13/13 [==============================] - 0s 29ms/step - loss: 0.7360\n",
            "13/13 [==============================] - 0s 30ms/step - loss: 2.1604\n",
            "CNN's Loss  0.43284574151039124\n",
            "VGG19's Loss  0.7360312938690186\n",
            "ResNet50's Loss  2.1603641510009766\n"
          ]
        }
      ],
      "source": [
        "################################## evaluate each model on test data #################################\n",
        "\n",
        "result_cnn  = cnn.evaluate(X_test, Y_test)\n",
        "result_VGG  = model_VGG.evaluate(X_test, Y_test)\n",
        "result_ResNet  = model_ResNet.evaluate(X_test, Y_test)\n",
        "\n",
        "print('CNN\\'s Loss ', result_cnn)\n",
        "print('VGG19\\'s Loss ', result_VGG)\n",
        "print('ResNet50\\'s Loss ', result_ResNet)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "liwFbfl8hRzG"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.11"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}