Day077_HW-訓練神經網路的細節與技巧.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Work\n",
    "1. 請將 Epoch 加到 500 個，並觀察 learning curve 的走勢\n",
    "2. 請將 Optimizer 換成 SGD，並觀察 learning curve 的走勢"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import keras\n",
    "\n",
    "# 本作業可以不需使用 GPU, 將 GPU 設定為 \"無\" (若有 GPU 且想開啟，可設為 \"0\")\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 從 Keras 的內建功能中，取得 train 與 test 資料集\n",
    "train, test = keras.datasets.cifar10.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 將 X 與 Y 獨立放進變數\n",
    "x_train, y_train = train\n",
    "x_test, y_test = test\n",
    "# 資料前處理 - 標準化\n",
    "x_train = x_train / 255.\n",
    "x_test = x_test / 255.\n",
    "\n",
    "# 將資料從圖形 (RGB) 轉為向量 (Single Vector)\n",
    "x_train = x_train.reshape((len(x_train), -1))\n",
    "x_test = x_test.reshape((len(x_test), -1))\n",
    "\n",
    "# 將目標轉為 one-hot encoding\n",
    "y_train = keras.utils.to_categorical(y_train, num_classes=10)\n",
    "y_test = keras.utils.to_categorical(y_test, num_classes=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_mlp():\n",
    "    \"\"\"Code Here\n",
    "    建立你的神經網路\n",
    "    \"\"\"\n",
    "\n",
    "    return model\n",
    "model = build_mlp()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Compile 模型\n",
    "\"\"\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "設定要訓練的 Epoch 數\n",
    "\"\"\"\n",
    "model.fit(x_train, y_train, \n",
    "          epochs=, \n",
    "          batch_size=256, \n",
    "          validation_data=(x_test, y_test), \n",
    "          shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "# 以視覺畫方式檢視訓練過程\n",
    "\n",
    "train_loss = model.history.history[\"loss\"]\n",
    "valid_loss = model.history.history[\"val_loss\"]\n",
    "\n",
    "train_acc = model.history.history[\"acc\"]\n",
    "valid_acc = model.history.history[\"val_acc\"]\n",
    "\n",
    "plt.plot(range(len(train_loss)), train_loss, label=\"train loss\")\n",
    "plt.plot(range(len(valid_loss)), valid_loss, label=\"valid loss\")\n",
    "plt.legend()\n",
    "plt.title(\"Loss\")\n",
    "plt.show()\n",
    "\n",
    "plt.plot(range(len(train_acc)), train_acc, label=\"train accuracy\")\n",
    "plt.plot(range(len(valid_acc)), valid_acc, label=\"valid accuracy\")\n",
    "plt.legend()\n",
    "plt.title(\"Accuracy\")\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}