Chapter1_Tensors/Variables_solution.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import print_function\n",
    "import torch\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.date(2017, 9, 28)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from datetime import date\n",
    "date.today()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "author = \"kyubyong. https://github.com/Kyubyong/pytorch_exercises\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.2.0_3'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'1.13.2'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "NOTE on notation\n",
    "\n",
    "    _x, _y, _z, ...: NumPy 0-d or 1-d arrays\n",
    "    _X, _Y, _Z, ...: NumPy 2-d or higer dimensional arrays\n",
    "    x, y, z, ...: 0-d or 1-d tensors\n",
    "    X, Y, Z, ...: 2-d or higher dimensional tensors\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.autograd import Variable"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Variables "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Q0. Create a variable `X` of the size (3, 2), filled with 1's."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Variable containing:\n",
      " 1  1\n",
      " 1  1\n",
      " 1  1\n",
      "[torch.FloatTensor of size 3x2]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "X = Variable(torch.ones(3, 2))\n",
    "print(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Q1. Get the tensor of Variable X."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " 0.8075  0.2259  0.7292\n",
      " 0.5840 -0.5985 -1.5452\n",
      " 0.7410 -0.2508 -1.1745\n",
      "[torch.FloatTensor of size 3x3]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "X = Variable(torch.randn(3, 3))\n",
    "Y = X.data\n",
    "print(Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Q2. Complete the code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2oAAAFpCAYAAADtINuMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa47b9e0898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a trainable variable `w` of scalar 10.\n",
    "w = Variable(torch.ones(1)*10, requires_grad=True)\n",
    "\n",
    "gs, ws, grads = [], [], []\n",
    "for i in range(10):\n",
    "    y = w ** 2 + 2\n",
    "    \n",
    "    # apply backpropagation to y.\n",
    "    y.backward()\n",
    "    \n",
    "    w.data -= 0.01 * w.grad.data\n",
    "    \n",
    "    gs.append(i)\n",
    "    ws.append(w.data[0])\n",
    "    grads.append(w.grad.data[0])\n",
    "\n",
    "plt.figure(figsize=(15,6))\n",
    "ax=plt.subplot(1, 2, 1)\n",
    "ax.scatter(gs, ws, c=\"b\", label=\"w\")\n",
    "ax.legend(loc=\"upper right\")\n",
    "ax=plt.subplot(1, 2, 2)\n",
    "plt.scatter(gs, grads, c=\"r\", label=\"gradient\")\n",
    "plt.legend(loc=\"upper left\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "Q3. Complete the code. <br>This is adapted from `http://pytorch.org/tutorials/beginner/examples_autograd/two_layer_net_autograd.html`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 010/500: loss 3622828.00000\n",
      "Epoch 020/500: loss 240678.50000\n",
      "Epoch 030/500: loss 68993.89062\n",
      "Epoch 040/500: loss 24136.43945\n",
      "Epoch 050/500: loss 9427.08691\n",
      "Epoch 060/500: loss 3968.63916\n",
      "Epoch 070/500: loss 1762.71545\n",
      "Epoch 080/500: loss 816.37415\n",
      "Epoch 090/500: loss 388.99170\n",
      "Epoch 100/500: loss 189.52725\n",
      "Epoch 110/500: loss 94.02654\n",
      "Epoch 120/500: loss 47.34197\n",
      "Epoch 130/500: loss 24.13431\n",
      "Epoch 140/500: loss 12.43430\n",
      "Epoch 150/500: loss 6.46430\n",
      "Epoch 160/500: loss 3.38794\n",
      "Epoch 170/500: loss 1.78757\n",
      "Epoch 180/500: loss 0.94889\n",
      "Epoch 190/500: loss 0.50667\n",
      "Epoch 200/500: loss 0.27179\n",
      "Epoch 210/500: loss 0.14644\n",
      "Epoch 220/500: loss 0.07919\n",
      "Epoch 230/500: loss 0.04301\n",
      "Epoch 240/500: loss 0.02346\n",
      "Epoch 250/500: loss 0.01288\n",
      "Epoch 260/500: loss 0.00713\n",
      "Epoch 270/500: loss 0.00402\n",
      "Epoch 280/500: loss 0.00232\n",
      "Epoch 290/500: loss 0.00139\n",
      "Epoch 300/500: loss 0.00086\n",
      "Epoch 310/500: loss 0.00056\n",
      "Epoch 320/500: loss 0.00038\n",
      "Epoch 330/500: loss 0.00026\n",
      "Epoch 340/500: loss 0.00019\n",
      "Epoch 350/500: loss 0.00014\n",
      "Epoch 360/500: loss 0.00011\n",
      "Epoch 370/500: loss 0.00009\n",
      "Epoch 380/500: loss 0.00007\n",
      "Epoch 390/500: loss 0.00006\n",
      "Epoch 400/500: loss 0.00005\n",
      "Epoch 410/500: loss 0.00004\n",
      "Epoch 420/500: loss 0.00003\n",
      "Epoch 430/500: loss 0.00003\n",
      "Epoch 440/500: loss 0.00003\n",
      "Epoch 450/500: loss 0.00002\n",
      "Epoch 460/500: loss 0.00002\n",
      "Epoch 470/500: loss 0.00002\n",
      "Epoch 480/500: loss 0.00002\n",
      "Epoch 490/500: loss 0.00001\n",
      "Epoch 500/500: loss 0.00001\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEDCAYAAADOc0QpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAFzpJREFUeJzt3X+QnVWd5/H39/7odCAJCokhEjAww47joKAbENcBWasc\ngXGkLJ1aUlPDjxWpmVVXS8stHavUmbLKGqmRVZmRpVZUthRxlJ1lhIFh1a3AlKM0mYRfEYyuLJ0J\npBMkIZAf/eO7f9ynm9s3ne5OcrtvP0/er6pbfe9zT9/7PaH59OnznPucyEwkSdVS63UBkqTuM9wl\nqYIMd0mqIMNdkirIcJekCjLcJamCehruEXFzRGyPiEdm0fb6iNhY3J6IiOfmo0ZJKqPo5Tr3iLgQ\n2APckplnHcb3fRB4fWb+xzkrTpJKrKcj98xcDzzbfiwifiMi7o6IByPivoh49RTfug64dV6KlKQS\navS6gCncBPxJZv48It4I/A3w1vEnI+JVwOnAD3tUnyQteAsq3CNiCfDvgL+NiPHDizqaXQ58NzNH\n57M2SSqTBRXutKaJnsvMc6Zpcznw/nmqR5JKaUEthczM3cD/jYg/BIiWs8efL+bfXw78uEclSlIp\n9Hop5K20gvq3ImIwIt4L/BHw3ojYBDwKXNb2LZcD304vZSlJ0+rpUkhJ0txYUNMykqTu6NkJ1eXL\nl+eaNWt69faSVEoPPvjgjsxcMVO7noX7mjVrGBgY6NXbS1IpRcSTs2nntIwkVZDhLkkVZLhLUgUt\ntE+oStJhGR4eZnBwkH379vW6lK7q7+9n9erVNJvNI/p+w11SqQ0ODrJ06VLWrFlD2zWpSi0z2blz\nJ4ODg5x++ulH9BpOy0gqtX379nHSSSdVJtgBIoKTTjrpqP4aMdwllV6Vgn3c0fapdOH++NPP81f/\n+Dg79uzvdSmStGCVLty3bN/Dl3+4hZ17DvS6FEkCYMmSJb0u4SClC/d6rfWnyuiYFzyTpEMpXbg3\nDHdJC1Rm8rGPfYyzzjqL1772tdx2220AbNu2jQsvvJBzzjmHs846i/vuu4/R0VGuuuqqibbXX399\nV2sp3VLIer0V7iNjYz2uRNJC8+d//yiP/evurr7ma165jE//we/Mqu3tt9/Oxo0b2bRpEzt27ODc\nc8/lwgsv5Fvf+hZvf/vb+eQnP8no6CgvvvgiGzduZOvWrTzyyCMAPPfcc12t25G7JHXJ/fffz7p1\n66jX66xcuZK3vOUtPPDAA5x77rl87Wtf4zOf+QwPP/wwS5cu5YwzzuCXv/wlH/zgB7n77rtZtmxZ\nV2sp38i9Nj5yN9wlTTbbEfZ8u/DCC1m/fj133nknV111FR/5yEe44oor2LRpE/fccw833ngj3/nO\nd7j55pu79p4lHLm3SnbkLmmhueCCC7jtttsYHR1laGiI9evXc9555/Hkk0+ycuVK3ve+93HNNdew\nYcMGduzYwdjYGO9+97v57Gc/y4YNG7paiyN3SeqSd73rXfz4xz/m7LPPJiL4/Oc/z8knn8w3vvEN\nrrvuOprNJkuWLOGWW25h69atXH311YwV5w8/97nPdbWW0ob7qCdUJS0Qe/bsAVqfKr3uuuu47rrr\nJj1/5ZVXcuWVVx70fd0erbcr4bRMMXIfdeQuSYdSunAfH7mPpeEuSYdSunBvOOcuqUNWcLB3tH0q\nXbh7+QFJ7fr7+9m5c2elAn78eu79/f1H/BqlO6E6vhTSOXdJAKtXr2ZwcJChoaFel9JV4zsxHanS\nhfv45QccuUsCaDabR7xbUZWVblrGOXdJmlnpwt117pI0s/KFezhyl6SZlC/cnXOXpBmVLty95K8k\nzax04e6FwyRpZjOGe0ScGhE/iojHIuLRiPjQFG0uiohdEbGxuH1qbsr1kr+SNBuzWec+Anw0MzdE\nxFLgwYi4NzMf62h3X2a+o/slTlYM3B25S9I0Zhy5Z+a2zNxQ3H8e2AycMteFHUpE0KiFSyElaRqH\nNeceEWuA1wM/meLpN0XEpoj4h4iYcq+riLg2IgYiYuBoPipcr4Ujd0maxqzDPSKWAN8DPpyZnduL\nbwBelZlnA18G/m6q18jMmzJzbWauXbFixZHW3Bq5e20ZSTqkWYV7RDRpBfs3M/P2zuczc3dm7inu\n3wU0I2J5VyttU3PkLknTms1qmQC+CmzOzC8cos3JRTsi4rzidXd2s9B2rTl3w12SDmU2q2XeDPwx\n8HBEbCyO/RlwGkBm3gi8B/jTiBgB9gKX5xxeXLleqzlyl6RpzBjumXk/EDO0uQG4oVtFzaRRC8YM\nd0k6pNJ9QhVcLSNJMylluDfqrnOXpOmUMtwduUvS9EoZ7q6WkaTplTLcXS0jSdMrabh7VUhJmk5J\nw92RuyRNp5Th7lUhJWl6pQz3uidUJWlapQx3V8tI0vRKGe59jRr7hp2WkaRDKWW4v/y4Pp594UCv\ny5CkBauU4X7i8Ya7JE2ntOG+d3iUvQdGe12KJC1IpQz35Uv6ANj5wv4eVyJJC1Mpw/3E4xcBODUj\nSYdQ0nBvjdzfecM/MYcbPklSaZUy3MenZQB27R3uYSWStDCVMtxPO/E43nj6iQBsf955d0nqVMpw\njwg+8rZ/A8Azu/f1uBpJWnhKGe4AK5f1A/DMbkfuktSptOH+imWtFTPbn3fkLkmdShvux/U1WNrf\nYLsjd0k6SGnDHWBZf5Pn9430ugxJWnBKHe6LmjX2j3gJAknqVO5wb9TZP+KlfyWp04zhHhGnRsSP\nIuKxiHg0Ij40RZuIiC9FxJaIeCgi3jA35U62qFEz3CVpCo1ZtBkBPpqZGyJiKfBgRNybmY+1tbkE\nOLO4vRH4SvF1Ti1q1Ng/7LSMJHWaceSemdsyc0Nx/3lgM3BKR7PLgFuy5Z+Bl0XEqq5X22FRs84+\nR+6SdJDDmnOPiDXA64GfdDx1CvBU2+NBDv4FQERcGxEDETEwNDR0eJVOod+RuyRNadbhHhFLgO8B\nH87M3UfyZpl5U2auzcy1K1asOJKXmGRRs84BR+6SdJBZhXtENGkF+zcz8/YpmmwFTm17vLo4Nqc8\noSpJU5vNapkAvgpszswvHKLZHcAVxaqZ84Fdmbmti3VOqRXuTstIUqfZrJZ5M/DHwMMRsbE49mfA\naQCZeSNwF3ApsAV4Ebi6+6UebFGjzv5hR+6S1GnGcM/M+4GYoU0C7+9WUbPV+oSq4S5JnUr+CdUa\nB0bHGBtzqz1JalfycK8DOHqXpA4lD/dW+Z5UlaTJyh3uzfFwd+QuSe1KHe7949MyrpiRpElKHe4v\njdydlpGkduUOd0+oStKUSh3ufZ5QlaQplTrcm7XWZ6uGR13nLkntyh3uxch9eNRpGUlqV+5wr7fK\nH3HkLkmTlDrcG8W0zAFH7pI0SanDffyEqiN3SZqs1OHemDih6shdktqVOtzH59wNd0marCLh7rSM\nJLUrebi3pmVGxhy5S1K7Uod7oxi5H/DyA5I0SanDvW98nbs7MUnSJKUO90YxLTPsyF2SJil3uLsU\nUpKmVOpwjwia9WDYaRlJmqTU4Q6t5ZBOy0jSZKUP90YtPKEqSR1KH+59jZoXDpOkDqUP92a9xojh\nLkmTzBjuEXFzRGyPiEcO8fxFEbErIjYWt091v8xDa9TDyw9IUofGLNp8HbgBuGWaNvdl5ju6UtFh\natZrLoWUpA4zjtwzcz3w7DzUckSaNcNdkjp1a879TRGxKSL+ISJ+51CNIuLaiBiIiIGhoaGuvHGz\nEW7WIUkduhHuG4BXZebZwJeBvztUw8y8KTPXZubaFStWdOGtoVFztYwkdTrqcM/M3Zm5p7h/F9CM\niOVHXdks9dVrjtwlqcNRh3tEnBwRUdw/r3jNnUf7urPVWi3jyF2S2s24WiYibgUuApZHxCDwaaAJ\nkJk3Au8B/jQiRoC9wOWZOW9D6Wa9xgsHRufr7SSpFGYM98xcN8PzN9BaKtkTzXp4bRlJ6lCJT6g6\nLSNJk5U+3Bv1mhcOk6QOpQ/3Zj3cQ1WSOpQ/3Gs1RsYMd0lqV/5wb3jhMEnqVP5w94SqJB3EcJek\nCqpAuHvhMEnqVPpwb9RaSyHn8UOxkrTglT7c+xqtLnhSVZJeUvpwb9QCwHl3SWpT+nBv1ltdcN5d\nkl5SgXBvjdzdsEOSXlKBcC9G7n5KVZImlD7cG0W4D484LSNJ40of7k7LSNLBKhDuTstIUqfKhLvT\nMpL0ktKHe6OYlhl25C5JE0of7n0TI3fDXZLGlT7cX5pzd1pGksaVPtwbrpaRpIOUPtz7vPyAJB2k\n9OE+cULVkbskTSh9uE8shTTcJWnCjOEeETdHxPaIeOQQz0dEfCkitkTEQxHxhu6XeWjNmtdzl6RO\nsxm5fx24eJrnLwHOLG7XAl85+rJmr9loTcuMOHKXpAkzhntmrgeenabJZcAt2fLPwMsiYlW3CpxJ\no+a0jCR16sac+ynAU22PB4tj82LiQ0xOy0jShHk9oRoR10bEQEQMDA0NdeU1XS0jSQfrRrhvBU5t\ne7y6OHaQzLwpM9dm5toVK1Z04a1dLSNJU+lGuN8BXFGsmjkf2JWZ27rwurPSnBi5Oy0jSeMaMzWI\niFuBi4DlETEIfBpoAmTmjcBdwKXAFuBF4Oq5KvYQ9dGohSN3SWozY7hn5roZnk/g/V2r6Ag06uGF\nwySpTek/oQqtefcDXvJXkiZUJtzdZk+SXlKRcA+32ZOkNhUJ95rb7ElSm+qEu0shJWlCRcI9vHCY\nJLWpRLg3ajXXuUtSm0qEe7PhtIwktatGuPsJVUmapBrhXq+5QbYktalEuDfqwQFH7pI0oRLh3lf3\nhKoktatEuDfq4bSMJLWpRLg3HblL0iTVCXcvPyBJEyoS7l44TJLaVSLcG17yV5ImqUS497lZhyRN\nUolwb7rNniRNUolwb7haRpImqUS4j1/PvbVXtySpGuFeCwCnZiSpUI1wb7S64dSMJLVUItwXFeG+\nf9hwlySoSLgvbtYB2Ds82uNKJGlhqEa49xnuktRuVuEeERdHxOMRsSUiPj7F81dFxFBEbCxu13S/\n1EPrL0bu+wx3SQKgMVODiKgDfw28DRgEHoiIOzLzsY6mt2XmB+agxhktNtwlaZLZjNzPA7Zk5i8z\n8wDwbeCyuS3r8IyP3Pce8ISqJMHswv0U4Km2x4PFsU7vjoiHIuK7EXHqVC8UEddGxEBEDAwNDR1B\nuVNz5C5Jk3XrhOrfA2sy83XAvcA3pmqUmTdl5trMXLtixYouvTUs7mt1wxOqktQym3DfCrSPxFcX\nxyZk5s7M3F88/O/Av+1OebPT71JISZpkNuH+AHBmRJweEX3A5cAd7Q0iYlXbw3cCm7tX4syclpGk\nyWZcLZOZIxHxAeAeoA7cnJmPRsRfAAOZeQfwnyPincAI8Cxw1RzWfBCXQkrSZDOGO0Bm3gXc1XHs\nU233PwF8orulzZ6rZSRpskp8QrVeC/oaNefcJalQiXCH1ry70zKS1FKZcO9v1gx3SSpUJtwXN+u8\neMBwlySoULgv6W+wZ/9Ir8uQpAWhMuF+wuImu/cO97oMSVoQKhPuy/qb7DLcJQmoWLjv3me4SxJU\nKdwXN9i91zl3SYIKhfsJi5vsHR7lwIifUpWkyoT7ssVNAKdmJIkqhXt/K9w9qSpJFQr3E8ZH7oa7\nJFUo3I9rhfuvXzzQ40okqfcqE+4nL+sHYNuufT2uRJJ6rzLh/oqli6gFPG24S1J1wr1Rr7FyWT//\n+pzhLkmVCXeAk0/o5+nde3tdhiT1XKXC/ZUnLHbkLklULNx/4xVLeHLnC+z1uu6SjnGVCvezXrmM\nsYTNT+/udSmS1FPVCvdTTgDg4cFdPa5EknqrUuG+6oR+TjvxOH74s+29LkWSeqpS4R4RXPLak/mn\nLTvY/rwnViUduyoV7gDrzj2NBP7qnid6XYok9Uzlwn3N8uO55oLTuW3gKf7y7p8xOpa9LkmS5t2s\nwj0iLo6IxyNiS0R8fIrnF0XEbcXzP4mINd0u9HB87Pd+i3XnncpX/s8vuOSL67n1p/+Poef397Ik\nSZpXkTn9yDYi6sATwNuAQeABYF1mPtbW5j8Br8vMP4mIy4F3ZeZ/mO51165dmwMDA0db/7TufGgb\nX/zBEzzxzB4i4Izlx/Pqk5fxm69Ywspl/axYuojlS/o4flGDxc06xy9qcFxfnUWNGhExp7VJ0pGI\niAczc+1M7RqzeK3zgC2Z+cvihb8NXAY81tbmMuAzxf3vAjdERORMvznm2O+/bhWXvvZkfvb08/xg\n8zM8NLiLh7fu4s6Ht834vfVaUI9ofe28FcfHtf8emHSfmOJYe9s46Nj4g6naSaqGy889lWsuOGNO\n32M24X4K8FTb40HgjYdqk5kjEbELOAnY0d4oIq4FrgU47bTTjrDkwxMR/PaqZfz2qmUTxw6MjLHz\nhf1s372fZ184wAsHRnjxwCgv7h/hxeFR9g2PMTaWjIwlY5mMjBZfx8YYHYPR4itA0vb7a4q77b/f\n2n/TjR+efCwPOoanDKTKWb5k0Zy/x2zCvWsy8ybgJmhNy8zne7fra9RYdcJiVp2wuFclSNKcms0J\n1a3AqW2PVxfHpmwTEQ3gBGBnNwqUJB2+2YT7A8CZEXF6RPQBlwN3dLS5A7iyuP8e4Ie9nm+XpGPZ\njNMyxRz6B4B7gDpwc2Y+GhF/AQxk5h3AV4H/ERFbgGdp/QKQJPXIrObcM/Mu4K6OY59qu78P+MPu\nliZJOlKV+4SqJMlwl6RKMtwlqYIMd0mqoBmvLTNnbxwxBDx5hN++nI5Pvx4D7POxwT4fG46mz6/K\nzBUzNepZuB+NiBiYzYVzqsQ+Hxvs87FhPvrstIwkVZDhLkkVVNZwv6nXBfSAfT422Odjw5z3uZRz\n7pKk6ZV15C5JmobhLkkVVLpwn2mz7rKKiJsjYntEPNJ27MSIuDcifl58fXlxPCLiS8W/wUMR8Ybe\nVX7kIuLUiPhRRDwWEY9GxIeK45Xtd0T0R8RPI2JT0ec/L46fXmwuv6XYbL6vOL6gNp8/UhFRj4h/\niYjvF48r3V+AiPhVRDwcERsjYqA4Nm8/26UK92Kz7r8GLgFeA6yLiNf0tqqu+TpwccexjwM/yMwz\ngR8Uj6HV/zOL27XAV+apxm4bAT6ama8BzgfeX/z3rHK/9wNvzcyzgXOAiyPifOAvgesz8zeBXwPv\nLdq/F/h1cfz6ol0ZfQjY3Pa46v0d9+8z85y2Ne3z97OdmaW5AW8C7ml7/AngE72uq4v9WwM80vb4\ncWBVcX8V8Hhx/78B66ZqV+Yb8L+Atx0r/QaOAzbQ2pN4B9Aojk/8nNPaR+FNxf1G0S56Xfth9nN1\nEWRvBb5Pa//3yva3rd+/ApZ3HJu3n+1SjdyZerPuU3pUy3xYmZnbivtPAyuL+5X7dyj+/H498BMq\n3u9iimIjsB24F/gF8FxmjhRN2vs1afN5YHzz+TL5r8B/AYpt5TmJavd3XAL/GBEPRsS1xbF5+9me\n1w2ydeQyMyOikutWI2IJ8D3gw5m5OyImnqtivzNzFDgnIl4G/E/g1T0uac5ExDuA7Zn5YERc1Ot6\n5tnvZubWiHgFcG9E/Kz9ybn+2S7byH02m3VXyTMRsQqg+Lq9OF6Zf4eIaNIK9m9m5u3F4cr3GyAz\nnwN+RGta4mXF5vIwuV9l33z+zcA7I+JXwLdpTc18ker2d0Jmbi2+bqf1S/w85vFnu2zhPpvNuquk\nfePxK2nNSY8fv6I4w34+sKvtT73SiNYQ/avA5sz8QttTle13RKwoRuxExGJa5xg20wr59xTNOvtc\n2s3nM/MTmbk6M9fQ+v/1h5n5R1S0v+Mi4viIWDp+H/g94BHm82e71ycdjuAkxaXAE7TmKT/Z63q6\n2K9bgW3AMK35tvfSmmv8AfBz4H8DJxZtg9aqoV8ADwNre13/Efb5d2nNSz4EbCxul1a538DrgH8p\n+vwI8Kni+BnAT4EtwN8Ci4rj/cXjLcXzZ/S6D0fR94uA7x8L/S36t6m4PTqeVfP5s+3lBySpgso2\nLSNJmgXDXZIqyHCXpAoy3CWpggx3Saogw12SKshwl6QK+v+QW+ovXoN90gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa47148ec18>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Untrainable variables\n",
    "# Create untrainable variables X and Y.\n",
    "X = Variable(torch.randn(64, 1000), requires_grad=False)\n",
    "Y = Variable(torch.randn(64, 10), requires_grad=False)\n",
    "\n",
    "# Trainable Variables\n",
    "w1 = Variable(torch.randn(1000, 100), requires_grad=True)\n",
    "w2 = Variable(torch.randn(100, 10), requires_grad=True)\n",
    "\n",
    "# Parameters\n",
    "n_epochs = 500\n",
    "learning_rate = 1e-6\n",
    "log_interval = 10\n",
    "\n",
    "losses = []\n",
    "for t in range(n_epochs):\n",
    "    Y_pred = X.matmul(w1).clamp(min=0).matmul(w2)\n",
    "    \n",
    "    # Apply L2 loss to Y pred and Y.\n",
    "    loss = (Y_pred - Y).pow(2).sum()\n",
    "    losses.append(loss.data[0])\n",
    "\n",
    "    loss.backward()\n",
    "    \n",
    "    w1.data -= learning_rate * w1.grad.data\n",
    "    w2.data -= learning_rate * w2.grad.data  \n",
    "\n",
    "    # Manually zero the gradients after updating weights\n",
    "    w1.grad.data.zero_()\n",
    "    w2.grad.data.zero_()\n",
    "    \n",
    "    if (t + 1) % log_interval == 0:\n",
    "        print(\"Epoch {:03d}/{:03d}: loss {:.5f}\".format(\n",
    "                t + 1, n_epochs, loss.data[0]))\n",
    "\n",
    "# plot\n",
    "plt.figure()\n",
    "plt.plot(losses, label=\"loss\")\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}