{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Step:  0 [k, b] [0.054201454, 0.10031855]\n",
      "Step:  100 [k, b] [0.10048652, 0.19973609]\n",
      "Step:  200 [k, b] [0.10003704, 0.1999799]\n",
      "Step:  300 [k, b] [0.10000283, 0.19999847]\n",
      "Step:  400 [k, b] [0.10000023, 0.19999987]\n",
      "Step:  500 [k, b] [0.10000023, 0.19999987]\n"
     ]
    }
   ],
   "source": [
    "x_data = np.random.rand(100)\n",
    "y_data = x_data * 0.1 + 0.2\n",
    "\n",
    "# 线性模型的斜率 & 偏置量\n",
    "k = tf.Variable(0.)\n",
    "b = tf.Variable(0.)\n",
    "y = k * x_data + b\n",
    "\n",
    "# 二次代价函数\n",
    "loss = tf.reduce_mean(tf.square(y_data - y))\n",
    "# 梯度下降\n",
    "optimizer = tf.train.GradientDescentOptimizer(0.2)\n",
    "# 最小化代价函数\n",
    "train = optimizer.minimize(loss)\n",
    "\n",
    "# 线性模型中，k, b作为变量，会在梯度下降法的作用下不断变化，以使得 loss函数越来越小\n",
    "with tf.Session() as sess:\n",
    "    tf.global_variables_initializer().run()\n",
    "    for step in range(501):\n",
    "        sess.run(train)\n",
    "        if step % 100 == 0:\n",
    "            print(\"Step: \", step, \"[k, b]\", sess.run([k, b]))"
   ]
  }
 ],
 "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}