01-2 取回你訓練好的神經網路.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 主題 01-2. 取回訓練好的神經網路\n",
    "\n",
    "【註】因 TensorFlow 2 已做了一些改變, 例如完全整合了 Keras。到 2021 年的今天, 有一些細節也做了調整。因此我們依新的規範修改了程式。最大的不同是, 以後大家直接安裝 tensorflow 即可, 不用再另外裝 keras。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. 找回我們的神經網路\n",
    "\n",
    "我們這裡需要把結構和 weights 取回, 要注意的是還需要再 compile 一次, 所以在 complie 中用到的 optimizer 等還是需要告訴 Keras 的。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import model_from_json\n",
    "from tensorflow.keras.optimizers import SGD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = model_from_json(open('handwriting_model_architecture.json').read())\n",
    "model.load_weights('handwriting_model_weights.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss='mse', optimizer=SGD(lr=0.1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. 資料再讀回來"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.datasets import mnist\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "我們只轉測試資料。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test = x_test.reshape(10000, 784)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. 找一個來試試"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# n_test = 1234\n",
    "\n",
    "n_test=9487"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = x_test[n_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fdb9439bf70>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.imshow(x.reshape(28,28), cmap='Greys')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(model.predict(x.reshape(1,784)), axis=-1)"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda env:tf2py38]",
   "language": "python",
   "name": "conda-env-tf2py38-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}