TensorflowTest3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "np.random.seed(0)\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Using some higher level tools to do training stuffs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "N, D, H = 64, 1000, 100\n",
    "x = tf.placeholder(tf.float32, shape=(N,D))\n",
    "y = tf.placeholder(tf.float32, shape=(N,D))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 1, Use Xavier initializer\n",
    "##### 2, `tf.layers` automatically sets up weight and bias for us!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "init = tf.contrib.layers.xavier_initializer()\n",
    "h = tf.layers.dense(inputs = x, units = H, \n",
    "                   activation = tf.nn.relu,\n",
    "                   kernel_initializer = init)\n",
    "y_pred = tf.layers.dense(inputs = h, units = D,\n",
    "                        kernel_initializer = init)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Define loss function, optimizer and updates:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "loss = tf.losses.mean_squared_error(y_pred, y)\n",
    "optimizer = tf.train.GradientDescentOptimizer(1.e0)\n",
    "updates = optimizer.minimize(loss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Do training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "t_list = []\n",
    "l_list = []\n",
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    values = {\n",
    "        x: np.random.randn(N,D),\n",
    "        y: np.random.randn(N,D),\n",
    "    }\n",
    "    for t in range(50):\n",
    "        loss_val, _ = sess.run([loss, updates], feed_dict = values)\n",
    "        t_list.append(t)\n",
    "        l_list.append(loss_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mpl\n",
    "%matplotlib inline \n",
    "\n",
    "mpl.rc('font',family='Times New Roman',size = 25)\n",
    "mpl.rcParams['xtick.major.size'] = 5\n",
    "mpl.rcParams['xtick.minor.size'] = 2\n",
    "mpl.rcParams['ytick.major.size'] = 5\n",
    "mpl.rcParams['ytick.minor.size'] = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA9MAAAGWCAYAAAB7B9CzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIhIS/8DgGymb04DrC4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x116466dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax  = plt.subplots(1,2,figsize=(16,6))\n",
    "ax[0].scatter(t_list, l_list)\n",
    "ax[0].set_xlabel(\"Number of steps\")\n",
    "ax[0].set_ylabel(\"Loss function\")\n",
    "ax[1].scatter(t_list, l_list)\n",
    "ax[1].set_xlabel(\"Number of steps\")\n",
    "#ax[1].set_ylabel(\"Loss function\")\n",
    "ax[1].set_yscale(\"log\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}