12-TrainPneumoniaFCN-Nadam.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "12-TrainPneumoniaFCN-Nadam.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "mount_file_id": "1KNQZDLntnWurZR4-nZY2HOPTb5gx8YFP",
      "authorship_tag": "ABX9TyOm5l7Eb/KXr8UFoqYWz1M4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "mRdIHAFo-tIf"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from time import time\n",
        "import random\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input, Flatten\n",
        "from keras.layers import Conv2D, SeparableConv2D\n",
        "from keras.layers import BatchNormalization, MaxPooling2D\n",
        "from keras.layers import Dense, Dropout"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "21QVu38WirVV"
      },
      "source": [
        "def load_numpy():\n",
        "    global X_train\n",
        "    global Y_train\n",
        "    global X_test \n",
        "    global Y_test  \n",
        "\n",
        "    with open('/content/drive/MyDrive/IDEC/chest-array.npy','rb') as f:\n",
        "        X_train = np.load(f)\n",
        "        Y_train = np.load(f)\n",
        "        X_test = np.load(f)\n",
        "        Y_test = np.load(f)\n",
        "\n",
        "    print(\"Read Numpy Data\")"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XBKqb7lE-QeH",
        "outputId": "74ac1244-3879-4178-f3e2-853bb3707e86"
      },
      "source": [
        "load_numpy()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Read Numpy Data\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qrd7qut82T1H"
      },
      "source": [
        "def conv_block(filters, inputs):\n",
        "    x = SeparableConv2D(filters,\n",
        "                        kernel_size=3,\n",
        "                        activation='relu',\n",
        "                        padding='same')(inputs)\n",
        "\n",
        "    x = SeparableConv2D(filters,\n",
        "                        kernel_size=3,\n",
        "                        activation='relu',\n",
        "                        padding='same')(x)\n",
        "\n",
        "    x = BatchNormalization()(x)\n",
        "\n",
        "    output = MaxPooling2D(pool_size=2)(x)\n",
        "\n",
        "    return output     "
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IuhE1qlG-3Lv"
      },
      "source": [
        "def dense_block(units, drop, inputs):\n",
        "    x = Dense(units=units,\n",
        "              activation='relu')(inputs)\n",
        "\n",
        "    x = BatchNormalization()(x)\n",
        "\n",
        "    output = Dropout(drop)(x)\n",
        "\n",
        "    return output"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xn5WVTcjQU2u"
      },
      "source": [
        "EPOCH = 200\n",
        "BATCH_SIZE = 100\n",
        "RES = 180\n",
        "SHAPE = (RES, RES, 3)"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F1OakI0R-4fH",
        "outputId": "b3246203-3209-42be-b54e-0cca7dd462b2"
      },
      "source": [
        "# Design Network\n",
        "\n",
        "input = Input(shape=(SHAPE))\n",
        "\n",
        "x = Conv2D(filters=16,\n",
        "           kernel_size=3,\n",
        "           activation='relu',\n",
        "           padding='same')(input)\n",
        "\n",
        "x = Conv2D(filters=16,\n",
        "           kernel_size=3,\n",
        "           activation='relu',\n",
        "           padding='same')(x)\n",
        "\n",
        "x = MaxPooling2D(pool_size=2)(x)\n",
        "\n",
        "x = conv_block(32,x)\n",
        "\n",
        "x = conv_block(64,x)\n",
        "\n",
        "x = conv_block(128,x)\n",
        "x = Dropout(0.2)(x)\n",
        "\n",
        "x = conv_block(256,x)\n",
        "x = Dropout(0.2)(x)\n",
        "\n",
        "x = Flatten()(x)\n",
        "\n",
        "x = dense_block(512, 0.7, x)\n",
        "x = dense_block(128, 0.5, x)\n",
        "x = dense_block(64, 0.3, x)\n",
        "\n",
        "output = Dense(units=1,\n",
        "          activation='sigmoid')(x)\n",
        "\n",
        "model = Model(inputs=input,\n",
        "              outputs=output)\n",
        "model.summary()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_1 (InputLayer)         [(None, 180, 180, 3)]     0         \n",
            "_________________________________________________________________\n",
            "conv2d (Conv2D)              (None, 180, 180, 16)      448       \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 180, 180, 16)      2320      \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 90, 90, 16)        0         \n",
            "_________________________________________________________________\n",
            "separable_conv2d (SeparableC (None, 90, 90, 32)        688       \n",
            "_________________________________________________________________\n",
            "separable_conv2d_1 (Separabl (None, 90, 90, 32)        1344      \n",
            "_________________________________________________________________\n",
            "batch_normalization (BatchNo (None, 90, 90, 32)        128       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 45, 45, 32)        0         \n",
            "_________________________________________________________________\n",
            "separable_conv2d_2 (Separabl (None, 45, 45, 64)        2400      \n",
            "_________________________________________________________________\n",
            "separable_conv2d_3 (Separabl (None, 45, 45, 64)        4736      \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 45, 45, 64)        256       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 22, 22, 64)        0         \n",
            "_________________________________________________________________\n",
            "separable_conv2d_4 (Separabl (None, 22, 22, 128)       8896      \n",
            "_________________________________________________________________\n",
            "separable_conv2d_5 (Separabl (None, 22, 22, 128)       17664     \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 22, 22, 128)       512       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_3 (MaxPooling2 (None, 11, 11, 128)       0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 11, 11, 128)       0         \n",
            "_________________________________________________________________\n",
            "separable_conv2d_6 (Separabl (None, 11, 11, 256)       34176     \n",
            "_________________________________________________________________\n",
            "separable_conv2d_7 (Separabl (None, 11, 11, 256)       68096     \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 11, 11, 256)       1024      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_4 (MaxPooling2 (None, 5, 5, 256)         0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 5, 5, 256)         0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 6400)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 512)               3277312   \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 512)               2048      \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 128)               65664     \n",
            "_________________________________________________________________\n",
            "batch_normalization_5 (Batch (None, 128)               512       \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "batch_normalization_6 (Batch (None, 64)                256       \n",
            "_________________________________________________________________\n",
            "dropout_4 (Dropout)          (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 1)                 65        \n",
            "=================================================================\n",
            "Total params: 3,496,801\n",
            "Trainable params: 3,494,433\n",
            "Non-trainable params: 2,368\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RQFCzYdN-4lx",
        "outputId": "8fc2c172-f269-400d-bc4a-6c8079472064"
      },
      "source": [
        "# Train Network\n",
        "\n",
        "model.compile(optimizer='nadam',\n",
        "              loss='binary_crossentropy',\n",
        "              metrics=['acc'])\n",
        "\n",
        "print(\"Start Train\\n\")\n",
        "\n",
        "begin = time()\n",
        "model.fit(x=X_train,\n",
        "          y=Y_train,\n",
        "          epochs=EPOCH,\n",
        "          batch_size=BATCH_SIZE,\n",
        "          verbose=1)\n",
        "end = time()\n",
        "\n",
        "print(\"Learning TIme : {:.2f}\".format(end-begin))\n",
        "\n",
        "model.save_weights('classification.h5')"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Start Train\n",
            "\n",
            "Epoch 1/200\n",
            "10/10 [==============================] - 12s 311ms/step - loss: 0.7395 - acc: 0.6325\n",
            "Epoch 2/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.3358 - acc: 0.8502\n",
            "Epoch 3/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.2596 - acc: 0.8871\n",
            "Epoch 4/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.1869 - acc: 0.9293\n",
            "Epoch 5/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.1609 - acc: 0.9380\n",
            "Epoch 6/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.1342 - acc: 0.9492\n",
            "Epoch 7/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.1077 - acc: 0.9589\n",
            "Epoch 8/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0948 - acc: 0.9587\n",
            "Epoch 9/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0985 - acc: 0.9686\n",
            "Epoch 10/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0885 - acc: 0.9677\n",
            "Epoch 11/200\n",
            "10/10 [==============================] - 3s 310ms/step - loss: 0.0606 - acc: 0.9758\n",
            "Epoch 12/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0573 - acc: 0.9797\n",
            "Epoch 13/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0476 - acc: 0.9852\n",
            "Epoch 14/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0379 - acc: 0.9914\n",
            "Epoch 15/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0386 - acc: 0.9891\n",
            "Epoch 16/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0326 - acc: 0.9844\n",
            "Epoch 17/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0425 - acc: 0.9842\n",
            "Epoch 18/200\n",
            "10/10 [==============================] - 3s 325ms/step - loss: 0.0337 - acc: 0.9862\n",
            "Epoch 19/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0256 - acc: 0.9949\n",
            "Epoch 20/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0167 - acc: 0.9951\n",
            "Epoch 21/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0232 - acc: 0.9939\n",
            "Epoch 22/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0197 - acc: 0.9969\n",
            "Epoch 23/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0292 - acc: 0.9866\n",
            "Epoch 24/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0221 - acc: 0.9922\n",
            "Epoch 25/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0184 - acc: 0.9944\n",
            "Epoch 26/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0160 - acc: 0.9964\n",
            "Epoch 27/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0156 - acc: 0.9927\n",
            "Epoch 28/200\n",
            "10/10 [==============================] - 3s 310ms/step - loss: 0.0200 - acc: 0.9919\n",
            "Epoch 29/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0128 - acc: 0.9971\n",
            "Epoch 30/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0069 - acc: 0.9997\n",
            "Epoch 31/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0132 - acc: 0.9969\n",
            "Epoch 32/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0063 - acc: 0.9997\n",
            "Epoch 33/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0123 - acc: 0.9952\n",
            "Epoch 34/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0264 - acc: 0.9868\n",
            "Epoch 35/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0576 - acc: 0.9773\n",
            "Epoch 36/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0201 - acc: 0.9935\n",
            "Epoch 37/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0193 - acc: 0.9912\n",
            "Epoch 38/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0120 - acc: 0.9945\n",
            "Epoch 39/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0104 - acc: 0.9967\n",
            "Epoch 40/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0090 - acc: 0.9968\n",
            "Epoch 41/200\n",
            "10/10 [==============================] - 3s 324ms/step - loss: 0.0076 - acc: 0.9983\n",
            "Epoch 42/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0088 - acc: 0.9964\n",
            "Epoch 43/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0135 - acc: 0.9935\n",
            "Epoch 44/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0092 - acc: 0.9967\n",
            "Epoch 45/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0101 - acc: 0.9975\n",
            "Epoch 46/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0207 - acc: 0.9946\n",
            "Epoch 47/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0131 - acc: 0.9965\n",
            "Epoch 48/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0064 - acc: 0.9995\n",
            "Epoch 49/200\n",
            "10/10 [==============================] - 3s 324ms/step - loss: 0.0147 - acc: 0.9941\n",
            "Epoch 50/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0200 - acc: 0.9939\n",
            "Epoch 51/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0102 - acc: 0.9962\n",
            "Epoch 52/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0145 - acc: 0.9960\n",
            "Epoch 53/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0083 - acc: 0.9962\n",
            "Epoch 54/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0056 - acc: 0.9981\n",
            "Epoch 55/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0068 - acc: 0.9969\n",
            "Epoch 56/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0069 - acc: 0.9972\n",
            "Epoch 57/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0100 - acc: 0.9978\n",
            "Epoch 58/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0049 - acc: 0.9987\n",
            "Epoch 59/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0033 - acc: 0.9997\n",
            "Epoch 60/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0037 - acc: 0.9982\n",
            "Epoch 61/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0113 - acc: 0.9972\n",
            "Epoch 62/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0030 - acc: 0.9998\n",
            "Epoch 63/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0072 - acc: 0.9965\n",
            "Epoch 64/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0044 - acc: 1.0000\n",
            "Epoch 65/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0020 - acc: 1.0000\n",
            "Epoch 66/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0018 - acc: 1.0000\n",
            "Epoch 67/200\n",
            "10/10 [==============================] - 3s 307ms/step - loss: 0.0143 - acc: 0.9941\n",
            "Epoch 68/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0042 - acc: 0.9986\n",
            "Epoch 69/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0090 - acc: 0.9958\n",
            "Epoch 70/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0290 - acc: 0.9908\n",
            "Epoch 71/200\n",
            "10/10 [==============================] - 3s 309ms/step - loss: 0.0143 - acc: 0.9982\n",
            "Epoch 72/200\n",
            "10/10 [==============================] - 3s 310ms/step - loss: 0.0028 - acc: 1.0000\n",
            "Epoch 73/200\n",
            "10/10 [==============================] - 3s 304ms/step - loss: 0.0049 - acc: 0.9982\n",
            "Epoch 74/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0026 - acc: 0.9989\n",
            "Epoch 75/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0030 - acc: 0.9992\n",
            "Epoch 76/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0075 - acc: 0.9978\n",
            "Epoch 77/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0117 - acc: 0.9946\n",
            "Epoch 78/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0126 - acc: 0.9945\n",
            "Epoch 79/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0106 - acc: 0.9937\n",
            "Epoch 80/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0051 - acc: 0.9990\n",
            "Epoch 81/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0050 - acc: 0.9993\n",
            "Epoch 82/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0360 - acc: 0.9883\n",
            "Epoch 83/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 0.0142 - acc: 0.9965\n",
            "Epoch 84/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 0.0105 - acc: 0.9973\n",
            "Epoch 85/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0120 - acc: 0.9952\n",
            "Epoch 86/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0186 - acc: 0.9905\n",
            "Epoch 87/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0380 - acc: 0.9892\n",
            "Epoch 88/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0197 - acc: 0.9956\n",
            "Epoch 89/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0072 - acc: 0.9955\n",
            "Epoch 90/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0056 - acc: 0.9985\n",
            "Epoch 91/200\n",
            "10/10 [==============================] - 3s 308ms/step - loss: 0.0140 - acc: 0.9933\n",
            "Epoch 92/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0209 - acc: 0.9887\n",
            "Epoch 93/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0143 - acc: 0.9924\n",
            "Epoch 94/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0101 - acc: 0.9957\n",
            "Epoch 95/200\n",
            "10/10 [==============================] - 3s 326ms/step - loss: 0.0465 - acc: 0.9782\n",
            "Epoch 96/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 0.0165 - acc: 0.9955\n",
            "Epoch 97/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0040 - acc: 0.9996\n",
            "Epoch 98/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0155 - acc: 0.9956\n",
            "Epoch 99/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0166 - acc: 0.9916\n",
            "Epoch 100/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0132 - acc: 0.9972\n",
            "Epoch 101/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0046 - acc: 0.9997\n",
            "Epoch 102/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0029 - acc: 1.0000\n",
            "Epoch 103/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0042 - acc: 0.9985\n",
            "Epoch 104/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0024 - acc: 0.9989\n",
            "Epoch 105/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0030 - acc: 0.9995\n",
            "Epoch 106/200\n",
            "10/10 [==============================] - 3s 309ms/step - loss: 0.0046 - acc: 0.9989\n",
            "Epoch 107/200\n",
            "10/10 [==============================] - 3s 324ms/step - loss: 0.0164 - acc: 0.9976\n",
            "Epoch 108/200\n",
            "10/10 [==============================] - 3s 324ms/step - loss: 0.0718 - acc: 0.9794\n",
            "Epoch 109/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0130 - acc: 0.9944\n",
            "Epoch 110/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0347 - acc: 0.9940\n",
            "Epoch 111/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0050 - acc: 0.9986\n",
            "Epoch 112/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0047 - acc: 0.9981\n",
            "Epoch 113/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0061 - acc: 0.9973\n",
            "Epoch 114/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0069 - acc: 0.9975\n",
            "Epoch 115/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0026 - acc: 0.9993\n",
            "Epoch 116/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0194 - acc: 0.9913\n",
            "Epoch 117/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0062 - acc: 0.9975\n",
            "Epoch 118/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0046 - acc: 0.9993\n",
            "Epoch 119/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0100 - acc: 0.9978\n",
            "Epoch 120/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0048 - acc: 0.9989\n",
            "Epoch 121/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0028 - acc: 0.9995\n",
            "Epoch 122/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0103 - acc: 0.9978\n",
            "Epoch 123/200\n",
            "10/10 [==============================] - 3s 308ms/step - loss: 0.0012 - acc: 1.0000\n",
            "Epoch 124/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0044 - acc: 0.9989\n",
            "Epoch 125/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0102 - acc: 0.9961\n",
            "Epoch 126/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0030 - acc: 0.9991\n",
            "Epoch 127/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0037 - acc: 0.9991\n",
            "Epoch 128/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 9.0012e-04 - acc: 0.9996\n",
            "Epoch 129/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0017 - acc: 1.0000\n",
            "Epoch 130/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0014 - acc: 1.0000\n",
            "Epoch 131/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0014 - acc: 0.9998\n",
            "Epoch 132/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 6.9488e-04 - acc: 0.9997\n",
            "Epoch 133/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 7.7240e-04 - acc: 1.0000\n",
            "Epoch 134/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0059 - acc: 0.9989\n",
            "Epoch 135/200\n",
            "10/10 [==============================] - 3s 322ms/step - loss: 0.0030 - acc: 0.9995\n",
            "Epoch 136/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0026 - acc: 0.9995\n",
            "Epoch 137/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0097 - acc: 0.9959\n",
            "Epoch 138/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0095 - acc: 0.9986\n",
            "Epoch 139/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0015 - acc: 0.9995\n",
            "Epoch 140/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0029 - acc: 0.9996\n",
            "Epoch 141/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0074 - acc: 0.9971\n",
            "Epoch 142/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0058 - acc: 0.9964\n",
            "Epoch 143/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0036 - acc: 0.9994\n",
            "Epoch 144/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0018 - acc: 0.9997\n",
            "Epoch 145/200\n",
            "10/10 [==============================] - 3s 324ms/step - loss: 0.0030 - acc: 0.9991\n",
            "Epoch 146/200\n",
            "10/10 [==============================] - 3s 309ms/step - loss: 9.8449e-04 - acc: 1.0000\n",
            "Epoch 147/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0031 - acc: 0.9993\n",
            "Epoch 148/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 4.5989e-04 - acc: 1.0000\n",
            "Epoch 149/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0019 - acc: 0.9995\n",
            "Epoch 150/200\n",
            "10/10 [==============================] - 3s 324ms/step - loss: 0.0030 - acc: 0.9989\n",
            "Epoch 151/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0212 - acc: 0.9943\n",
            "Epoch 152/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0066 - acc: 0.9976\n",
            "Epoch 153/200\n",
            "10/10 [==============================] - 3s 325ms/step - loss: 0.0020 - acc: 0.9989\n",
            "Epoch 154/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0022 - acc: 0.9989\n",
            "Epoch 155/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0074 - acc: 0.9978\n",
            "Epoch 156/200\n",
            "10/10 [==============================] - 3s 326ms/step - loss: 0.0039 - acc: 0.9995\n",
            "Epoch 157/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0061 - acc: 0.9993\n",
            "Epoch 158/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0011 - acc: 0.9996\n",
            "Epoch 159/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0049 - acc: 0.9955\n",
            "Epoch 160/200\n",
            "10/10 [==============================] - 3s 325ms/step - loss: 6.3805e-04 - acc: 1.0000\n",
            "Epoch 161/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 8.3110e-04 - acc: 1.0000\n",
            "Epoch 162/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0066 - acc: 0.9979\n",
            "Epoch 163/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0018 - acc: 0.9991\n",
            "Epoch 164/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0024 - acc: 0.9992\n",
            "Epoch 165/200\n",
            "10/10 [==============================] - 3s 314ms/step - loss: 0.0014 - acc: 0.9998\n",
            "Epoch 166/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0288 - acc: 0.9925\n",
            "Epoch 167/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 0.0029 - acc: 0.9992\n",
            "Epoch 168/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0102 - acc: 0.9960\n",
            "Epoch 169/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0022 - acc: 0.9993\n",
            "Epoch 170/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0091 - acc: 0.9949\n",
            "Epoch 171/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 7.2328e-04 - acc: 1.0000\n",
            "Epoch 172/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 7.1561e-04 - acc: 1.0000\n",
            "Epoch 173/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0064 - acc: 0.9971\n",
            "Epoch 174/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 6.9445e-04 - acc: 0.9998\n",
            "Epoch 175/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0035 - acc: 0.9987\n",
            "Epoch 176/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0056 - acc: 0.9982\n",
            "Epoch 177/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 5.1026e-04 - acc: 1.0000\n",
            "Epoch 178/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0079 - acc: 0.9982\n",
            "Epoch 179/200\n",
            "10/10 [==============================] - 3s 307ms/step - loss: 6.4910e-04 - acc: 1.0000\n",
            "Epoch 180/200\n",
            "10/10 [==============================] - 3s 323ms/step - loss: 0.0075 - acc: 0.9973\n",
            "Epoch 181/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0012 - acc: 0.9997\n",
            "Epoch 182/200\n",
            "10/10 [==============================] - 3s 310ms/step - loss: 0.0061 - acc: 0.9977\n",
            "Epoch 183/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0016 - acc: 0.9993\n",
            "Epoch 184/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0016 - acc: 0.9997\n",
            "Epoch 185/200\n",
            "10/10 [==============================] - 3s 325ms/step - loss: 0.0033 - acc: 0.9996\n",
            "Epoch 186/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 4.6089e-04 - acc: 1.0000\n",
            "Epoch 187/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 6.4784e-04 - acc: 1.0000\n",
            "Epoch 188/200\n",
            "10/10 [==============================] - 3s 312ms/step - loss: 0.0072 - acc: 0.9968\n",
            "Epoch 189/200\n",
            "10/10 [==============================] - 3s 315ms/step - loss: 0.0027 - acc: 0.9985\n",
            "Epoch 190/200\n",
            "10/10 [==============================] - 3s 319ms/step - loss: 0.0010 - acc: 1.0000\n",
            "Epoch 191/200\n",
            "10/10 [==============================] - 3s 311ms/step - loss: 0.0014 - acc: 0.9995\n",
            "Epoch 192/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 9.9871e-04 - acc: 1.0000\n",
            "Epoch 193/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 0.0027 - acc: 0.9980\n",
            "Epoch 194/200\n",
            "10/10 [==============================] - 3s 321ms/step - loss: 0.0011 - acc: 0.9995\n",
            "Epoch 195/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0049 - acc: 0.9972\n",
            "Epoch 196/200\n",
            "10/10 [==============================] - 3s 316ms/step - loss: 0.0032 - acc: 0.9986\n",
            "Epoch 197/200\n",
            "10/10 [==============================] - 3s 318ms/step - loss: 0.0013 - acc: 0.9995\n",
            "Epoch 198/200\n",
            "10/10 [==============================] - 3s 320ms/step - loss: 0.0012 - acc: 1.0000\n",
            "Epoch 199/200\n",
            "10/10 [==============================] - 3s 313ms/step - loss: 7.9228e-04 - acc: 1.0000\n",
            "Epoch 200/200\n",
            "10/10 [==============================] - 3s 317ms/step - loss: 7.4054e-04 - acc: 1.0000\n",
            "Learning TIme : 643.99\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jvd88S5jtgRA",
        "outputId": "b95be121-2fbf-49e2-dc8e-147d56e394ed"
      },
      "source": [
        "score = model.evaluate(x=X_test,\n",
        "                       y=Y_test,\n",
        "                       verbose=1)\n",
        "print(\"Final Loss : {:.2f}\".format(score[0]))\n",
        "print(\"Final Accuracy : {:.2f}\".format(score[1]))"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7/7 [==============================] - 1s 41ms/step - loss: 1.6821 - acc: 0.7500\n",
            "Final Loss : 1.68\n",
            "Final Accuracy : 0.75\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 323
        },
        "id": "YuCuShalvhFA",
        "outputId": "ab9e2c3c-31f2-466b-f914-dee099ada6bd"
      },
      "source": [
        "total = len(X_test)\n",
        "pick = random.sample(range(0,total),1)\n",
        "print(pick)\n",
        "\n",
        "plt.imshow(X_test[pick].reshape(RES, RES, 3))\n",
        "img = X_test[pick].reshape(1, RES, RES, 3)\n",
        "pred = model.predict(img)\n",
        "\n",
        "print(\"Label :\", Y_test[pick])\n",
        "print(\"Pred :\", pred)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[196]\n",
            "Label : [1.]\n",
            "Pred : [[0.9999975]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}