2.2_MNIST_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 卷積神經網路 CNN "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 28, 28, 16)        416       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 14, 14, 16)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 14, 14, 32)        12832     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 7, 7, 32)          0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 7, 7, 32)          0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 1568)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 128)               200832    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 215,370\n",
      "Trainable params: 215,370\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/10\n",
      "375/375 - 16s - loss: 0.3977 - accuracy: 0.8727 - val_loss: 0.0787 - val_accuracy: 0.9768\n",
      "Epoch 2/10\n",
      "375/375 - 16s - loss: 0.1341 - accuracy: 0.9583 - val_loss: 0.0578 - val_accuracy: 0.9825\n",
      "Epoch 3/10\n",
      "375/375 - 16s - loss: 0.1022 - accuracy: 0.9686 - val_loss: 0.0474 - val_accuracy: 0.9868\n",
      "Epoch 4/10\n",
      "375/375 - 16s - loss: 0.0859 - accuracy: 0.9747 - val_loss: 0.0433 - val_accuracy: 0.9882\n",
      "Epoch 5/10\n",
      "375/375 - 16s - loss: 0.0744 - accuracy: 0.9767 - val_loss: 0.0374 - val_accuracy: 0.9895\n",
      "Epoch 6/10\n",
      "375/375 - 17s - loss: 0.0676 - accuracy: 0.9793 - val_loss: 0.0369 - val_accuracy: 0.9903\n",
      "Epoch 7/10\n",
      "375/375 - 16s - loss: 0.0609 - accuracy: 0.9815 - val_loss: 0.0336 - val_accuracy: 0.9905\n",
      "Epoch 8/10\n",
      "375/375 - 16s - loss: 0.0549 - accuracy: 0.9830 - val_loss: 0.0336 - val_accuracy: 0.9913\n",
      "Epoch 9/10\n",
      "375/375 - 16s - loss: 0.0534 - accuracy: 0.9834 - val_loss: 0.0311 - val_accuracy: 0.9911\n",
      "Epoch 10/10\n",
      "375/375 - 16s - loss: 0.0481 - accuracy: 0.9856 - val_loss: 0.0341 - val_accuracy: 0.9906\n",
      "\n",
      "Testing ...\n",
      "訓練資料集的準確度 = 0.99\n",
      "測試資料集的準確度 = 0.99\n",
      "Saving Model: mnist.h5 ...\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from tensorflow.keras.datasets import mnist\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import Flatten\n",
    "from tensorflow.keras.layers import Conv2D\n",
    "from tensorflow.keras.layers import MaxPooling2D\n",
    "from tensorflow.keras.layers import Dropout\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "\n",
    "# 指定亂數種子\n",
    "seed = 7\n",
    "np.random.seed(seed)\n",
    "# 載入資料集\n",
    "(X_train, Y_train), (X_test, Y_test) = mnist.load_data()\n",
    "\n",
    "# 將圖片轉換成 4D 張量\n",
    "X_train = X_train.reshape(X_train.shape[0], 28, 28, 1).astype(\"float32\")\n",
    "X_test = X_test.reshape(X_test.shape[0], 28, 28, 1).astype(\"float32\")\n",
    "\n",
    "# 因為是固定範圍, 所以執行正規化, 從 0-255 至 0-1\n",
    "X_train = X_train / 255\n",
    "X_test = X_test / 255\n",
    "\n",
    "# One-hot編碼\n",
    "Y_train = to_categorical(Y_train)\n",
    "Y_test = to_categorical(Y_test)\n",
    "\n",
    "\n",
    "# 定義模型\n",
    "model = Sequential()\n",
    "model.add(Conv2D(16, kernel_size=(5, 5), padding=\"same\",\n",
    "                 input_shape=(28, 28, 1), activation=\"relu\"))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Conv2D(32, kernel_size=(5, 5), padding=\"same\",\n",
    "                 activation=\"relu\"))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128, activation=\"relu\"))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(10, activation=\"softmax\"))\n",
    "model.summary()   # 顯示模型摘要資訊\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# 編譯模型\n",
    "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\",\n",
    "              metrics=[\"accuracy\"])\n",
    "# 訓練模型\n",
    "history = model.fit(X_train, Y_train, validation_split=0.2,\n",
    "                    epochs=10, batch_size=128, verbose=2)\n",
    "\n",
    "# 評估模型\n",
    "print(\"\\nTesting ...\")\n",
    "loss, accuracy = model.evaluate(X_train, Y_train, verbose=0)\n",
    "print(\"訓練資料集的準確度 = {:.2f}\".format(accuracy))\n",
    "loss, accuracy = model.evaluate(X_test, Y_test, verbose=0)\n",
    "print(\"測試資料集的準確度 = {:.2f}\".format(accuracy))\n",
    "# 儲存Keras模型\n",
    "print(\"Saving Model: mnist.h5 ...\")\n",
    "\n",
    "model.save(\"mnist_CNN.h5\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "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"
    },
    {
     "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": [
    "# 顯示圖表來分析模型的訓練過程\n",
    "import matplotlib.pyplot as plt\n",
    "# 顯示訓練和驗證損失\n",
    "loss = history.history[\"loss\"]\n",
    "epochs = range(1, len(loss)+1)\n",
    "val_loss = history.history[\"val_loss\"]\n",
    "plt.plot(epochs, loss, \"bo-\", label=\"Training Loss\")\n",
    "plt.plot(epochs, val_loss, \"ro--\", label=\"Validation Loss\")\n",
    "plt.title(\"Training and Validation Loss\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "# 顯示訓練和驗證準確度\n",
    "acc = history.history[\"acc\"]\n",
    "epochs = range(1, len(acc)+1)\n",
    "val_acc = history.history[\"val_acc\"]\n",
    "plt.plot(epochs, acc, \"bo-\", label=\"Training Acc\")\n",
    "plt.plot(epochs, val_acc, \"ro--\", label=\"Validation Acc\")\n",
    "plt.title(\"Training and Validation Accuracy\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}