tf_keras_CNN_MNIST.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.7.4"
    },
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jjyjung/ai/blob/gh-pages/tf_keras_CNN_MNIST.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "\n",
        "drive.mount('/content/gdrive')   # google drive를 google colab에 연결. 처음 실행 시, 인증 필요"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "E--gXv3WnY6a",
        "outputId": "6203aaab-e3fc-4fbe-a81e-635462a4542a"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qTEQyMB4hu2e"
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras import layers, utils, Sequential\n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "from tensorflow.keras.datasets import mnist"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EU-u6jREtQtz"
      },
      "source": [
        "# hyperparameters (실험변경 편의를 위하여 상단에 기재하였음)\n",
        "batch_size = 100\n",
        "epochs = 10\n",
        "learning_rate = 0.1\n",
        "dropout_rate = 0.1\n",
        "patience = 5"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QA2jvpXkREOY",
        "outputId": "e63cf101-9b11-4252-d4a7-12cb54dc138e"
      },
      "source": [
        "# input image dimensions, output dimension\n",
        "img_rows, img_cols = 28, 28\n",
        "num_classes = 10\n",
        "\n",
        "# the data, shuffled and split between train and test sets\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11490434/11490434 [==============================] - 2s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VWlFawQbhu2g",
        "outputId": "9cfd12b4-fc73-4000-cb6d-e9a7fd428be1"
      },
      "source": [
        "# the data, shuffled and split between train and test sets\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
        "\n",
        "x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)  # 이미지 개수, 행, 열, 흑백(1채널)\n",
        "x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)\n",
        "input_shape = (img_rows, img_cols, 1)\n",
        "\n",
        "x_train = x_train.astype('float32')\n",
        "x_test = x_test.astype('float32')\n",
        "x_train /= 255\n",
        "x_test /= 255\n",
        "print('x_train shape:', x_train.shape)\n",
        "print(x_train.shape[0], 'train samples')\n",
        "print(x_test.shape[0], 'test samples')\n",
        "\n",
        "# convert class vectors to binary class matrices\n",
        "y_train = utils.to_categorical(y_train, num_classes)\n",
        "y_test = utils.to_categorical(y_test, num_classes)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "x_train shape: (60000, 28, 28, 1)\n",
            "60000 train samples\n",
            "10000 test samples\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V_ZPXPOPhu2g",
        "outputId": "78b3d3c1-4b4b-4392-ae4d-d1b45592f358"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(layers.Conv2D(32, kernel_size=(3, 3),\n",
        "                 activation='relu',\n",
        "                 input_shape=input_shape))\n",
        "model.add(layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
        "model.add(layers.Dropout(dropout_rate))\n",
        "\n",
        "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\n",
        "model.add(layers.Dropout(dropout_rate))\n",
        "model.add(layers.Flatten())\n",
        "\n",
        "model.add(layers.Dense(32, activation='relu'))\n",
        "model.add(layers.Dropout(dropout_rate))\n",
        "model.add(layers.Dense(num_classes, activation='softmax'))\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 26, 26, 32)        320       \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2  (None, 13, 13, 32)        0         \n",
            " D)                                                              \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 13, 13, 32)        0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 11, 11, 64)        18496     \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPoolin  (None, 5, 5, 64)          0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 5, 5, 64)          0         \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 1600)              0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 32)                51232     \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 32)                0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 70378 (274.91 KB)\n",
            "Trainable params: 70378 (274.91 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yEb5iRHUv8sT",
        "outputId": "aababfb1-6866-4b21-91fe-b0eef3bb200c"
      },
      "source": [
        "# compile model\n",
        "model.compile(\n",
        "    loss = 'categorical_crossentropy',\n",
        "    optimizer = 'Adam',\n",
        "    metrics = ['accuracy'])\n",
        "\n",
        "# train model\n",
        "early_stop = EarlyStopping(monitor = 'val_loss', patience = patience, mode='min')\n",
        "hist = model.fit(x_train, y_train,\n",
        "          batch_size = batch_size,\n",
        "          epochs = epochs,\n",
        "          validation_split = 0.2,\n",
        "          callbacks = [early_stop],\n",
        "          shuffle = True,\n",
        "          verbose = 1)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "480/480 [==============================] - 14s 6ms/step - loss: 0.3459 - accuracy: 0.8919 - val_loss: 0.0913 - val_accuracy: 0.9738\n",
            "Epoch 2/10\n",
            "480/480 [==============================] - 3s 6ms/step - loss: 0.1074 - accuracy: 0.9669 - val_loss: 0.0627 - val_accuracy: 0.9814\n",
            "Epoch 3/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0784 - accuracy: 0.9764 - val_loss: 0.0481 - val_accuracy: 0.9851\n",
            "Epoch 4/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0637 - accuracy: 0.9803 - val_loss: 0.0491 - val_accuracy: 0.9854\n",
            "Epoch 5/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0521 - accuracy: 0.9834 - val_loss: 0.0494 - val_accuracy: 0.9854\n",
            "Epoch 6/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0483 - accuracy: 0.9853 - val_loss: 0.0390 - val_accuracy: 0.9885\n",
            "Epoch 7/10\n",
            "480/480 [==============================] - 3s 7ms/step - loss: 0.0407 - accuracy: 0.9873 - val_loss: 0.0367 - val_accuracy: 0.9893\n",
            "Epoch 8/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0367 - accuracy: 0.9889 - val_loss: 0.0354 - val_accuracy: 0.9898\n",
            "Epoch 9/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0332 - accuracy: 0.9889 - val_loss: 0.0348 - val_accuracy: 0.9898\n",
            "Epoch 10/10\n",
            "480/480 [==============================] - 2s 5ms/step - loss: 0.0298 - accuracy: 0.9906 - val_loss: 0.0331 - val_accuracy: 0.9912\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KeM75AIT0HWz"
      },
      "source": [
        "\"Epoch 3/10\"의 의미는?\n",
        "\n",
        "\"38s 77ms/step\"의 의미는?\n",
        "\n",
        "\"loss: 0.1105 - accuracy: 0.9666 - val_loss: 0.0624 - val_accuracy: 0.9827\"의 의미는?"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "fig, loss_ax = plt.subplots()\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(hist.history['loss'], 'y', label='train loss')\n",
        "loss_ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
        "acc_ax.plot(hist.history['accuracy'], 'b', label='train acc')\n",
        "acc_ax.plot(hist.history['val_accuracy'], 'g', label='val acc')\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "toZUlxyXaJA9",
        "outputId": "24d79089-6e79-4461-c272-34a899e0435a"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1PtSrMe-hu2h",
        "outputId": "2944e6ca-27f2-49a9-acc1-cc7d570e08b9"
      },
      "source": [
        "# evaluate model\n",
        "score = model.evaluate(x_test, y_test, verbose = 1)\n",
        "print('test_loss:', round(score[0], 4))\n",
        "print('test_acc:', round(score[1], 4))"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.0282 - accuracy: 0.9908\n",
            "test_loss: 0.0282\n",
            "test_acc: 0.9908\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dir = \"/content/gdrive/My Drive/Colab Notebooks/ai/CNN/\"  ### [중요!!] Colab의 경우, 파일을 저장할 위치를 기록하시오.\n",
        "model.save(dir + \"cnn_mnist.h5\")\n",
        "model.save(dir + \"cnn_mnist.keras\")"
      ],
      "metadata": {
        "id": "6sBk_qODGK_s"
      },
      "execution_count": 11,
      "outputs": []
    }
  ]
}