├── README.md ├── gapqa-architecture-1.png ├── gapqa-architecture.pdf ├── Task1-Understanding-Backpropagation ├── Neural-Network-1.png ├── README.md ├── Neural-Network-2.svg └── Understanding-Backpropagation.ipynb ├── Task4-Tensorflow-Intro ├── README.md ├── Tensorflow-Beginner-Practice.ipynb └── Tensorflow-Softmax-Classifier-Toy-Data.ipynb ├── Task2-Implementing-Backpropagation └── README.md └── Task3-Vanishing-Gradients └── README.md /README.md: -------------------------------------------------------------------------------- 1 | # learning-deep-learning 2 | My practical approach to learning Neural Network concepts 3 | -------------------------------------------------------------------------------- /gapqa-architecture-1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/rakeshchada/learning-deep-learning/HEAD/gapqa-architecture-1.png -------------------------------------------------------------------------------- /gapqa-architecture.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/rakeshchada/learning-deep-learning/HEAD/gapqa-architecture.pdf -------------------------------------------------------------------------------- /Task1-Understanding-Backpropagation/Neural-Network-1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/rakeshchada/learning-deep-learning/HEAD/Task1-Understanding-Backpropagation/Neural-Network-1.png -------------------------------------------------------------------------------- /Task4-Tensorflow-Intro/README.md: -------------------------------------------------------------------------------- 1 | # Getting familiar with Tensorflow basics 2 | In this task, we will play around with tensorflow for a bit. We will first implement a simple Linear Regression model. 3 | Finally, we will implement the Softmax model. 4 | -------------------------------------------------------------------------------- /Task1-Understanding-Backpropagation/README.md: -------------------------------------------------------------------------------- 1 | # Understanding Backpropagation 2 | In this task, we will go through the basics of neural networks. We will then formulate a simple task of representing mathematical functions via Neural Networks. We will use this as a framework for understanding backpropagation concepts. 3 | -------------------------------------------------------------------------------- /Task2-Implementing-Backpropagation/README.md: -------------------------------------------------------------------------------- 1 | # Implementing Backpropagation 2 | In this task, we will implement softmax and a single layer neural network from scratch. This will include implementing both the forward and backward passes of the training. 3 | We will train these classifiers on a toy dataset and compare the decision boundaries. 4 | -------------------------------------------------------------------------------- /Task3-Vanishing-Gradients/README.md: -------------------------------------------------------------------------------- 1 | # Vanishing Gradients Visualization 2 | In this notebook, we will work through the vanishing gradients problem through several visualizations. We will look at a few ways to address this problem and understand them through visualizations. 3 | We will also build a general multi-layer neural network to help understand the effect of depth of the network on its performance. 4 | -------------------------------------------------------------------------------- /Task1-Understanding-Backpropagation/Neural-Network-2.svg: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | background 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | Layer 1 17 | 18 | 19 | 20 | 21 | 22 | y 23 | x 24 | sig 25 | + 26 | 27 | 28 | 29 | + 30 | 31 | 32 | 33 | sig 34 | 35 | ^2 36 | 37 | 38 | + 39 | 40 | 41 | 42 | 1/x 43 | 44 | 45 | * 46 | 47 | 48 | 49 | f 50 | 51 | 52 | -------------------------------------------------------------------------------- /Task1-Understanding-Backpropagation/Understanding-Backpropagation.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "### Credit to http://cs231n.github.io/optimization-2/" 8 | ] 9 | }, 10 | { 11 | "cell_type": "markdown", 12 | "metadata": {}, 13 | "source": [ 14 | "### Imports" 15 | ] 16 | }, 17 | { 18 | "cell_type": "code", 19 | "execution_count": 3, 20 | "metadata": { 21 | "collapsed": true 22 | }, 23 | "outputs": [], 24 | "source": [ 25 | "%matplotlib inline\n", 26 | "from matplotlib import pyplot as plt\n", 27 | "import numpy as np\n", 28 | "import math\n", 29 | "from IPython.display import Image, SVG\n", 30 | "plt.rcParams['figure.figsize'] = (10.0, 8.0) # set default size of plots\n", 31 | "plt.rcParams['image.interpolation'] = 'nearest'\n", 32 | "plt.rcParams['image.cmap'] = 'gray'" 33 | ] 34 | }, 35 | { 36 | "cell_type": "markdown", 37 | "metadata": {}, 38 | "source": [ 39 | "### A simple neural network - representing sigmoid function (a.k.a Logistic regression)" 40 | ] 41 | }, 42 | { 43 | "cell_type": "markdown", 44 | "metadata": {}, 45 | "source": [ 46 | "![title](Neural-Network-1.png)" 47 | ] 48 | }, 49 | { 50 | "cell_type": "markdown", 51 | "metadata": {}, 52 | "source": [ 53 | "### The function that the above neural network represents" 54 | ] 55 | }, 56 | { 57 | "cell_type": "markdown", 58 | "metadata": {}, 59 | "source": [ 60 | "# $f(w,x) = \\frac{1}{1+e^{-(w_0x_0 + w_1x_1 + w_2)}}$ #" 61 | ] 62 | }, 63 | { 64 | "cell_type": "markdown", 65 | "metadata": {}, 66 | "source": [ 67 | "### Forwardprop and Backprop through sigmoid" 68 | ] 69 | }, 70 | { 71 | "cell_type": "code", 72 | "execution_count": 2, 73 | "metadata": { 74 | "collapsed": false 75 | }, 76 | "outputs": [ 77 | { 78 | "name": "stdout", 79 | "output_type": "stream", 80 | "text": [ 81 | "-0.534446645389 -0.534446645389 -0.196611933241 0.196611933241 0.196611933241 0.196611933241 0.196611933241\n", 82 | "-0.196611933241 -0.393223866483 0.393223866483 -0.589835799724 0.196611933241\n" 83 | ] 84 | } 85 | ], 86 | "source": [ 87 | "W = [2,-3,-3] # assume some random weights and data\n", 88 | "X = [-1, -2]\n", 89 | "\n", 90 | "# Forward pass\n", 91 | "dot = W[0]*X[0] + W[1]*X[1]\n", 92 | "dotplusw2 = dot + W[2]\n", 93 | "minus_dot = (-1) * dotplusw2\n", 94 | "exp_dot = math.exp(minus_dot)\n", 95 | "f = 1.0 / (1 + exp_dot) # sigmoid function\n", 96 | "\n", 97 | "# Backprop\n", 98 | "df = 1\n", 99 | "dplusone = (-1)*(f*f)\n", 100 | "dexp = dplusone\n", 101 | "dminusone = exp_dot * dexp\n", 102 | "ddotplusw2 = (-1) * dminusone\n", 103 | "dw2 = ddotplusw2\n", 104 | "ddot = ddotplusw2\n", 105 | "dw0x0 = ddot\n", 106 | "dw1x1 = ddot\n", 107 | "dw0 = X[0] * dw0x0\n", 108 | "dx0 = W[0] * dw0x0\n", 109 | "dw1 = X[1] * dw1x1\n", 110 | "dx1 = W[1] * dw1x1\n", 111 | "\n", 112 | "print dplusone, dexp, dminusone, ddotplusw2, ddot, dw0x0, dw1x1\n", 113 | "\n", 114 | "print dw0, dw1, dx0, dx1, dw2" 115 | ] 116 | }, 117 | { 118 | "cell_type": "markdown", 119 | "metadata": {}, 120 | "source": [ 121 | "# $\\sigma(x) = \\frac{1}{1+e^{-x}}$#\n", 122 | "# $\\rightarrow \\hspace{0.3in} \\frac{d\\sigma(x)}{dx} = \\frac{e^{-x}}{(1+e^{-x})^2} = \\left( \\frac{1 + e^{-x} - 1}{1 + e^{-x}} \\right) \\left( \\frac{1}{1+e^{-x}} \\right) = \\left( 1 - \\sigma(x) \\right) \\sigma(x)$ #" 123 | ] 124 | }, 125 | { 126 | "cell_type": "markdown", 127 | "metadata": {}, 128 | "source": [ 129 | "So one could essentially avoid all the individual sigmoid computations and use the above formula" 130 | ] 131 | }, 132 | { 133 | "cell_type": "code", 134 | "execution_count": 3, 135 | "metadata": { 136 | "collapsed": false 137 | }, 138 | "outputs": [ 139 | { 140 | "data": { 141 | "text/plain": [ 142 | "0.19661193324148185" 143 | ] 144 | }, 145 | "execution_count": 3, 146 | "metadata": {}, 147 | "output_type": "execute_result" 148 | } 149 | ], 150 | "source": [ 151 | "(1-f) * f" 152 | ] 153 | }, 154 | { 155 | "cell_type": "code", 156 | "execution_count": 4, 157 | "metadata": { 158 | "collapsed": false 159 | }, 160 | "outputs": [ 161 | { 162 | "data": { 163 | "text/plain": [ 164 | "-2.7755575615628914e-17" 165 | ] 166 | }, 167 | "execution_count": 4, 168 | "metadata": {}, 169 | "output_type": "execute_result" 170 | } 171 | ], 172 | "source": [ 173 | "((1-f) * f) - ddotplusw2" 174 | ] 175 | }, 176 | { 177 | "cell_type": "markdown", 178 | "metadata": {}, 179 | "source": [ 180 | "### A slightly complex function" 181 | ] 182 | }, 183 | { 184 | "cell_type": "markdown", 185 | "metadata": {}, 186 | "source": [ 187 | "# $f(x,y) = \\frac{x + \\sigma(y)}{\\sigma(x) + (x+y)^2}$ #" 188 | ] 189 | }, 190 | { 191 | "cell_type": "markdown", 192 | "metadata": {}, 193 | "source": [ 194 | "### Neural network for the above function (excuse the drawing :D)" 195 | ] 196 | }, 197 | { 198 | "cell_type": "code", 199 | "execution_count": 4, 200 | "metadata": { 201 | "collapsed": false 202 | }, 203 | "outputs": [ 204 | { 205 | "data": { 206 | "image/svg+xml": [ 207 | "\n", 208 | " \n", 209 | " \n", 210 | " background\n", 211 | " \n", 212 | " \n", 213 | " \n", 214 | " \n", 215 | " \n", 216 | " \n", 217 | " \n", 218 | " \n", 219 | " \n", 220 | " \n", 221 | " \n", 222 | " Layer 1\n", 223 | " \n", 224 | " \n", 225 | " \n", 226 | " \n", 227 | " \n", 228 | " y\n", 229 | " x\n", 230 | " sig\n", 231 | " +\n", 232 | " \n", 233 | " \n", 234 | " \n", 235 | " +\n", 236 | " \n", 237 | " \n", 238 | " \n", 239 | " sig\n", 240 | " \n", 241 | " ^2\n", 242 | " \n", 243 | " \n", 244 | " +\n", 245 | " \n", 246 | " \n", 247 | " \n", 248 | " 1/x\n", 249 | " \n", 250 | " \n", 251 | " *\n", 252 | " \n", 253 | " \n", 254 | " \n", 255 | " f\n", 256 | " \n", 257 | " \n", 258 | "" 259 | ], 260 | "text/plain": [ 261 | "" 262 | ] 263 | }, 264 | "execution_count": 4, 265 | "metadata": {}, 266 | "output_type": "execute_result" 267 | } 268 | ], 269 | "source": [ 270 | "SVG(\"Neural-Network-2.svg\")" 271 | ] 272 | }, 273 | { 274 | "cell_type": "markdown", 275 | "metadata": {}, 276 | "source": [ 277 | "### Forward and backprop through the above neural network" 278 | ] 279 | }, 280 | { 281 | "cell_type": "code", 282 | "execution_count": 5, 283 | "metadata": { 284 | "collapsed": false 285 | }, 286 | "outputs": [ 287 | { 288 | "name": "stdout", 289 | "output_type": "stream", 290 | "text": [ 291 | "2.05956979557 1.59223275148\n" 292 | ] 293 | } 294 | ], 295 | "source": [ 296 | "x = 3 # example values\n", 297 | "y = -4\n", 298 | "\n", 299 | "# forward pass\n", 300 | "sigy = 1.0 / (1 + math.exp(-y)) # sigmoid in numerator #(1)\n", 301 | "num = x + sigy # numerator #(2)\n", 302 | "sigx = 1.0 / (1 + math.exp(-x)) # sigmoid in denominator #(3)\n", 303 | "xpy = x + y #(4)\n", 304 | "xpysqr = xpy**2 #(5)\n", 305 | "den = sigx + xpysqr # denominator #(6)\n", 306 | "invden = 1.0 / den #(7)\n", 307 | "f = num * invden # done! #(8)\n", 308 | "\n", 309 | "# backprop f = num * invden\n", 310 | "dnum = invden # gradient on numerator #(8)\n", 311 | "dinvden = num #(8)\n", 312 | "# backprop invden = 1.0 / den \n", 313 | "dden = (-1.0 / (den**2)) * dinvden #(7)\n", 314 | "# backprop den = sigx + xpysqr\n", 315 | "dsigx = (1) * dden #(6)\n", 316 | "dxpysqr = (1) * dden #(6)\n", 317 | "# backprop xpysqr = xpy**2\n", 318 | "dxpy = (2 * xpy) * dxpysqr #(5)\n", 319 | "# backprop xpy = x + y\n", 320 | "dx = (1) * dxpy #(4)\n", 321 | "dy = (1) * dxpy #(4)\n", 322 | "# backprop sigx = 1.0 / (1 + math.exp(-x))\n", 323 | "dx += ((1 - sigx) * sigx) * dsigx # Gradients add up at forks #(3)\n", 324 | "# backprop num = x + sigy\n", 325 | "dx += (1) * dnum # Gradients add up at forks #(2)\n", 326 | "dsigy = (1) * dnum #(2)\n", 327 | "# backprop sigy = 1.0 / (1 + math.exp(-y))\n", 328 | "dy += ((1 - sigy) * sigy) * dsigy # Gradients add up at forks\n", 329 | "\n", 330 | "print dx, dy" 331 | ] 332 | } 333 | ], 334 | "metadata": { 335 | "kernelspec": { 336 | "display_name": "Python 2", 337 | "language": "python", 338 | "name": "python2" 339 | }, 340 | "language_info": { 341 | "codemirror_mode": { 342 | "name": "ipython", 343 | "version": 2 344 | }, 345 | "file_extension": ".py", 346 | "mimetype": "text/x-python", 347 | "name": "python", 348 | "nbconvert_exporter": "python", 349 | "pygments_lexer": "ipython2", 350 | "version": "2.7.10" 351 | } 352 | }, 353 | "nbformat": 4, 354 | "nbformat_minor": 1 355 | } 356 | -------------------------------------------------------------------------------- /Task4-Tensorflow-Intro/Tensorflow-Beginner-Practice.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "### Testing out basics" 8 | ] 9 | }, 10 | { 11 | "cell_type": "code", 12 | "execution_count": 1, 13 | "metadata": { 14 | "collapsed": true 15 | }, 16 | "outputs": [], 17 | "source": [ 18 | "import tensorflow as tf" 19 | ] 20 | }, 21 | { 22 | "cell_type": "code", 23 | "execution_count": 2, 24 | "metadata": { 25 | "collapsed": false 26 | }, 27 | "outputs": [], 28 | "source": [ 29 | "interactive_session = tf.InteractiveSession()" 30 | ] 31 | }, 32 | { 33 | "cell_type": "code", 34 | "execution_count": 3, 35 | "metadata": { 36 | "collapsed": true 37 | }, 38 | "outputs": [], 39 | "source": [ 40 | "a = tf.zeros((2,2))\n", 41 | "b = tf.ones((2,2))" 42 | ] 43 | }, 44 | { 45 | "cell_type": "code", 46 | "execution_count": 4, 47 | "metadata": { 48 | "collapsed": false 49 | }, 50 | "outputs": [ 51 | { 52 | "data": { 53 | "text/plain": [ 54 | "" 55 | ] 56 | }, 57 | "execution_count": 4, 58 | "metadata": {}, 59 | "output_type": "execute_result" 60 | } 61 | ], 62 | "source": [ 63 | "tf.reduce_sum(b, reduction_indices=1)" 64 | ] 65 | }, 66 | { 67 | "cell_type": "code", 68 | "execution_count": 5, 69 | "metadata": { 70 | "collapsed": false 71 | }, 72 | "outputs": [ 73 | { 74 | "data": { 75 | "text/plain": [ 76 | "array([ 2., 2.], dtype=float32)" 77 | ] 78 | }, 79 | "execution_count": 5, 80 | "metadata": {}, 81 | "output_type": "execute_result" 82 | } 83 | ], 84 | "source": [ 85 | "tf.reduce_sum(b, reduction_indices=1).eval()" 86 | ] 87 | }, 88 | { 89 | "cell_type": "code", 90 | "execution_count": 6, 91 | "metadata": { 92 | "collapsed": false 93 | }, 94 | "outputs": [ 95 | { 96 | "data": { 97 | "text/plain": [ 98 | "TensorShape([Dimension(2), Dimension(2)])" 99 | ] 100 | }, 101 | "execution_count": 6, 102 | "metadata": {}, 103 | "output_type": "execute_result" 104 | } 105 | ], 106 | "source": [ 107 | "a.get_shape()" 108 | ] 109 | }, 110 | { 111 | "cell_type": "code", 112 | "execution_count": 7, 113 | "metadata": { 114 | "collapsed": false 115 | }, 116 | "outputs": [ 117 | { 118 | "data": { 119 | "text/plain": [ 120 | "array([[ 0., 0.],\n", 121 | " [ 0., 0.]], dtype=float32)" 122 | ] 123 | }, 124 | "execution_count": 7, 125 | "metadata": {}, 126 | "output_type": "execute_result" 127 | } 128 | ], 129 | "source": [ 130 | "a.eval()" 131 | ] 132 | }, 133 | { 134 | "cell_type": "code", 135 | "execution_count": 8, 136 | "metadata": { 137 | "collapsed": false 138 | }, 139 | "outputs": [ 140 | { 141 | "data": { 142 | "text/plain": [ 143 | "array([[ 0., 0., 0., 0.]], dtype=float32)" 144 | ] 145 | }, 146 | "execution_count": 8, 147 | "metadata": {}, 148 | "output_type": "execute_result" 149 | } 150 | ], 151 | "source": [ 152 | "tf.reshape(a, (1,4)).eval()" 153 | ] 154 | }, 155 | { 156 | "cell_type": "code", 157 | "execution_count": 9, 158 | "metadata": { 159 | "collapsed": false 160 | }, 161 | "outputs": [ 162 | { 163 | "data": { 164 | "text/plain": [ 165 | "array([[ 0.],\n", 166 | " [ 0.],\n", 167 | " [ 0.],\n", 168 | " [ 0.]], dtype=float32)" 169 | ] 170 | }, 171 | "execution_count": 9, 172 | "metadata": {}, 173 | "output_type": "execute_result" 174 | } 175 | ], 176 | "source": [ 177 | "tf.reshape(a, (4,1)).eval()" 178 | ] 179 | }, 180 | { 181 | "cell_type": "code", 182 | "execution_count": 10, 183 | "metadata": { 184 | "collapsed": false 185 | }, 186 | "outputs": [ 187 | { 188 | "data": { 189 | "text/plain": [ 190 | "array([[ 0., 0.],\n", 191 | " [ 0., 0.]], dtype=float32)" 192 | ] 193 | }, 194 | "execution_count": 10, 195 | "metadata": {}, 196 | "output_type": "execute_result" 197 | } 198 | ], 199 | "source": [ 200 | "tf.matmul(a,b).eval()" 201 | ] 202 | }, 203 | { 204 | "cell_type": "code", 205 | "execution_count": 11, 206 | "metadata": { 207 | "collapsed": false 208 | }, 209 | "outputs": [ 210 | { 211 | "data": { 212 | "text/plain": [ 213 | "" 214 | ] 215 | }, 216 | "execution_count": 11, 217 | "metadata": {}, 218 | "output_type": "execute_result" 219 | } 220 | ], 221 | "source": [ 222 | "b" 223 | ] 224 | }, 225 | { 226 | "cell_type": "code", 227 | "execution_count": 12, 228 | "metadata": { 229 | "collapsed": false 230 | }, 231 | "outputs": [ 232 | { 233 | "name": "stdout", 234 | "output_type": "stream", 235 | "text": [ 236 | "Tensor(\"ones:0\", shape=(2, 2), dtype=float32)\n" 237 | ] 238 | } 239 | ], 240 | "source": [ 241 | "print(b)" 242 | ] 243 | }, 244 | { 245 | "cell_type": "code", 246 | "execution_count": 13, 247 | "metadata": { 248 | "collapsed": true 249 | }, 250 | "outputs": [], 251 | "source": [ 252 | "weights = tf.Variable(tf.zeros((2,2)), name=\"weights\")" 253 | ] 254 | }, 255 | { 256 | "cell_type": "code", 257 | "execution_count": 14, 258 | "metadata": { 259 | "collapsed": false 260 | }, 261 | "outputs": [], 262 | "source": [ 263 | "#weights.eval() - throws exception due to uninitialization" 264 | ] 265 | }, 266 | { 267 | "cell_type": "code", 268 | "execution_count": 15, 269 | "metadata": { 270 | "collapsed": false 271 | }, 272 | "outputs": [ 273 | { 274 | "data": { 275 | "text/plain": [ 276 | "" 277 | ] 278 | }, 279 | "execution_count": 15, 280 | "metadata": {}, 281 | "output_type": "execute_result" 282 | } 283 | ], 284 | "source": [ 285 | "tf.initialize_all_variables()" 286 | ] 287 | }, 288 | { 289 | "cell_type": "code", 290 | "execution_count": 16, 291 | "metadata": { 292 | "collapsed": false 293 | }, 294 | "outputs": [ 295 | { 296 | "name": "stdout", 297 | "output_type": "stream", 298 | "text": [ 299 | "[[ 0. 0.]\n", 300 | " [ 0. 0.]]\n" 301 | ] 302 | } 303 | ], 304 | "source": [ 305 | "with tf.Session() as sess:\n", 306 | " sess.run(tf.initialize_all_variables())\n", 307 | " weights.eval()\n", 308 | " print sess.run(weights)" 309 | ] 310 | }, 311 | { 312 | "cell_type": "code", 313 | "execution_count": 17, 314 | "metadata": { 315 | "collapsed": false 316 | }, 317 | "outputs": [ 318 | { 319 | "name": "stdout", 320 | "output_type": "stream", 321 | "text": [ 322 | "1\n", 323 | "2\n", 324 | "3\n" 325 | ] 326 | } 327 | ], 328 | "source": [ 329 | "d = tf.Variable(0, name=\"Counter\")\n", 330 | "add_one = tf.add(d,1)\n", 331 | "update_d = tf.assign(d, add_one)\n", 332 | "with tf.Session() as sess:\n", 333 | " sess.run(tf.initialize_all_variables())\n", 334 | " for i in range(3):\n", 335 | " sess.run(update_d)\n", 336 | " print(sess.run(d))" 337 | ] 338 | }, 339 | { 340 | "cell_type": "code", 341 | "execution_count": 18, 342 | "metadata": { 343 | "collapsed": false 344 | }, 345 | "outputs": [ 346 | { 347 | "data": { 348 | "text/plain": [ 349 | "" 350 | ] 351 | }, 352 | "execution_count": 18, 353 | "metadata": {}, 354 | "output_type": "execute_result" 355 | } 356 | ], 357 | "source": [ 358 | "add_one" 359 | ] 360 | }, 361 | { 362 | "cell_type": "markdown", 363 | "metadata": {}, 364 | "source": [ 365 | "### Linear Regression" 366 | ] 367 | }, 368 | { 369 | "cell_type": "code", 370 | "execution_count": 19, 371 | "metadata": { 372 | "collapsed": true 373 | }, 374 | "outputs": [], 375 | "source": [ 376 | "import numpy as np\n", 377 | "%matplotlib inline" 378 | ] 379 | }, 380 | { 381 | "cell_type": "code", 382 | "execution_count": 20, 383 | "metadata": { 384 | "collapsed": true 385 | }, 386 | "outputs": [], 387 | "source": [ 388 | "from matplotlib import pyplot as plt" 389 | ] 390 | }, 391 | { 392 | "cell_type": "code", 393 | "execution_count": 21, 394 | "metadata": { 395 | "collapsed": false 396 | }, 397 | "outputs": [], 398 | "source": [ 399 | "import seaborn" 400 | ] 401 | }, 402 | { 403 | "cell_type": "code", 404 | "execution_count": 22, 405 | "metadata": { 406 | "collapsed": false 407 | }, 408 | "outputs": [ 409 | { 410 | "data": { 411 | "text/plain": [ 412 | "" 413 | ] 414 | }, 415 | "execution_count": 22, 416 | "metadata": {}, 417 | "output_type": "execute_result" 418 | }, 419 | { 420 | "data": { 421 | "image/png": "iVBORw0KGgoAAAANSUhEUgAAArQAAAHcCAYAAAAncTlwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\njUILAAAAo1FoAQAAYDQKLQAAAIz2/wpf7DLMSLlZAAAAAElFTkSuQmCC\n", 422 | "text/plain": [ 423 | "" 424 | ] 425 | }, 426 | "metadata": {}, 427 | "output_type": "display_data" 428 | } 429 | ], 430 | "source": [ 431 | "x = np.arange(0,100,0.1)\n", 432 | "y = x + 20 * np.sin(x/10)\n", 433 | "plt.scatter(x, y)" 434 | ] 435 | }, 436 | { 437 | "cell_type": "code", 438 | "execution_count": 23, 439 | "metadata": { 440 | "collapsed": true 441 | }, 442 | "outputs": [], 443 | "source": [ 444 | "n_samples = 1000\n", 445 | "batch_size = 100" 446 | ] 447 | }, 448 | { 449 | "cell_type": "code", 450 | "execution_count": 24, 451 | "metadata": { 452 | "collapsed": true 453 | }, 454 | "outputs": [], 455 | "source": [ 456 | "x_data = np.reshape(x, (n_samples,1))\n", 457 | "y_data = np.reshape(y, (n_samples, 1))" 458 | ] 459 | }, 460 | { 461 | "cell_type": "code", 462 | "execution_count": 25, 463 | "metadata": { 464 | "collapsed": true 465 | }, 466 | "outputs": [], 467 | "source": [ 468 | "x_input = tf.placeholder(tf.float32, shape=(batch_size, 1))\n", 469 | "y_input = tf.placeholder(tf.float32, shape=(batch_size, 1))" 470 | ] 471 | }, 472 | { 473 | "cell_type": "code", 474 | "execution_count": 26, 475 | "metadata": { 476 | "collapsed": false 477 | }, 478 | "outputs": [], 479 | "source": [ 480 | "with tf.variable_scope(\"linear_regression\"):\n", 481 | " W = tf.get_variable(\"weights\", (1,1), initializer=tf.random_normal_initializer())\n", 482 | " b = tf.get_variable(\"bias\", (1,), initializer=tf.constant_initializer(0.0))\n", 483 | " y_pred = tf.matmul(x_input,W) + b\n", 484 | " loss = tf.reduce_sum((y - y_pred)**2)/n_samples" 485 | ] 486 | }, 487 | { 488 | "cell_type": "code", 489 | "execution_count": 27, 490 | "metadata": { 491 | "collapsed": true 492 | }, 493 | "outputs": [], 494 | "source": [ 495 | "opt = tf.train.AdamOptimizer()\n", 496 | "opt_operation = opt.minimize(loss)" 497 | ] 498 | }, 499 | { 500 | "cell_type": "code", 501 | "execution_count": 28, 502 | "metadata": { 503 | "collapsed": true 504 | }, 505 | "outputs": [], 506 | "source": [ 507 | "if 'session' in locals() and interactive_session is not None:\n", 508 | " print('Close interactive session')\n", 509 | " interactive_session.close()" 510 | ] 511 | }, 512 | { 513 | "cell_type": "code", 514 | "execution_count": 29, 515 | "metadata": { 516 | "collapsed": false 517 | }, 518 | "outputs": [ 519 | { 520 | "name": "stdout", 521 | "output_type": "stream", 522 | "text": [ 523 | "2.25614e+06\n", 524 | "2.43183e+06\n", 525 | "2.31179e+06\n", 526 | "2.16133e+06\n", 527 | "2.54426e+06\n", 528 | "2.35845e+06\n", 529 | "2.39975e+06\n", 530 | "2.1677e+06\n", 531 | "2.25646e+06\n", 532 | "2.42176e+06\n", 533 | "2.37617e+06\n", 534 | "2.43084e+06\n", 535 | "2.60132e+06\n", 536 | "2.44399e+06\n", 537 | "2.36708e+06\n", 538 | "2.22893e+06\n", 539 | "2.1983e+06\n", 540 | "2.29019e+06\n", 541 | "2.32304e+06\n", 542 | "2.34377e+06\n", 543 | "2.4913e+06\n", 544 | "2.31421e+06\n", 545 | "2.31149e+06\n", 546 | "2.26452e+06\n", 547 | "2.54935e+06\n", 548 | "2.23671e+06\n", 549 | "2.35057e+06\n", 550 | "2.43478e+06\n", 551 | "2.205e+06\n", 552 | "2.22286e+06\n", 553 | "2.1635e+06\n", 554 | "2.28358e+06\n", 555 | "2.47716e+06\n", 556 | "1.98313e+06\n", 557 | "2.46643e+06\n", 558 | "2.18862e+06\n", 559 | "2.15147e+06\n", 560 | "2.40125e+06\n", 561 | "2.63574e+06\n", 562 | "2.39022e+06\n", 563 | "2.23217e+06\n", 564 | "2.29198e+06\n", 565 | "2.16732e+06\n", 566 | "2.50349e+06\n", 567 | "2.10509e+06\n", 568 | "2.38082e+06\n", 569 | "2.35742e+06\n", 570 | "2.19042e+06\n", 571 | "2.09911e+06\n", 572 | "2.17594e+06\n", 573 | "2.37786e+06\n", 574 | "2.24528e+06\n", 575 | "2.25904e+06\n", 576 | "1.99264e+06\n", 577 | "2.31908e+06\n", 578 | "2.25577e+06\n", 579 | "2.14531e+06\n", 580 | "2.12874e+06\n", 581 | "2.42401e+06\n", 582 | "2.32386e+06\n", 583 | "2.37008e+06\n", 584 | "2.23923e+06\n", 585 | "2.25091e+06\n", 586 | "2.41755e+06\n", 587 | "2.11321e+06\n", 588 | "2.13969e+06\n", 589 | "2.28334e+06\n", 590 | "2.43771e+06\n", 591 | "2.40326e+06\n", 592 | "2.09848e+06\n", 593 | "2.18583e+06\n", 594 | "2.16114e+06\n", 595 | "2.37793e+06\n", 596 | "2.22876e+06\n", 597 | "2.18223e+06\n", 598 | "2.25421e+06\n", 599 | "2.30217e+06\n", 600 | "2.36085e+06\n", 601 | "2.50623e+06\n", 602 | "2.21176e+06\n", 603 | "2.14829e+06\n", 604 | "2.43469e+06\n", 605 | "1.98888e+06\n", 606 | "2.40416e+06\n", 607 | "2.23203e+06\n", 608 | "2.27157e+06\n", 609 | "2.27616e+06\n", 610 | "2.20703e+06\n", 611 | "2.29333e+06\n", 612 | "2.21749e+06\n", 613 | "2.58785e+06\n", 614 | "2.27777e+06\n", 615 | "2.36099e+06\n", 616 | "2.44879e+06\n", 617 | "2.07239e+06\n", 618 | "2.09787e+06\n", 619 | "2.06752e+06\n", 620 | "2.0778e+06\n", 621 | "2.0859e+06\n", 622 | "2.05987e+06\n", 623 | "2.17779e+06\n", 624 | "2.22998e+06\n", 625 | "2.31492e+06\n", 626 | "2.1862e+06\n", 627 | "2.24129e+06\n", 628 | "2.05166e+06\n", 629 | "2.13432e+06\n", 630 | "2.39354e+06\n", 631 | "2.10333e+06\n", 632 | "1.89049e+06\n", 633 | "2.1422e+06\n", 634 | "2.18068e+06\n", 635 | "2.19646e+06\n", 636 | "2.29694e+06\n", 637 | "2.20959e+06\n", 638 | "2.26419e+06\n", 639 | "2.53305e+06\n", 640 | "2.24068e+06\n", 641 | "2.1364e+06\n", 642 | "1.96638e+06\n", 643 | "2.24959e+06\n", 644 | "2.01117e+06\n", 645 | "2.52258e+06\n", 646 | "2.01816e+06\n", 647 | "2.22046e+06\n", 648 | "2.11363e+06\n", 649 | "2.11237e+06\n", 650 | "2.20121e+06\n", 651 | "2.23086e+06\n", 652 | "1.99477e+06\n", 653 | "2.24162e+06\n", 654 | "1.93353e+06\n", 655 | "2.25909e+06\n", 656 | "2.31406e+06\n", 657 | "2.14398e+06\n", 658 | "2.1406e+06\n", 659 | "2.22138e+06\n", 660 | "1.93684e+06\n", 661 | "2.05599e+06\n", 662 | "1.98393e+06\n", 663 | "2.20562e+06\n", 664 | "2.00847e+06\n", 665 | "2.39587e+06\n", 666 | "2.09508e+06\n", 667 | "2.02461e+06\n", 668 | "2.1225e+06\n", 669 | "2.26899e+06\n", 670 | "2.25132e+06\n", 671 | "2.24356e+06\n", 672 | "1.94763e+06\n", 673 | "2.15038e+06\n", 674 | "2.18772e+06\n", 675 | "2.14355e+06\n", 676 | "2.10506e+06\n", 677 | "2.17178e+06\n", 678 | "2.19175e+06\n", 679 | "2.19775e+06\n", 680 | "2.1847e+06\n", 681 | "2.21731e+06\n", 682 | "2.05375e+06\n", 683 | "2.06044e+06\n", 684 | "2.05295e+06\n", 685 | "2.04229e+06\n", 686 | "2.31391e+06\n", 687 | "1.97896e+06\n", 688 | "2.01516e+06\n", 689 | "1.89489e+06\n", 690 | "1.99276e+06\n", 691 | "1.96978e+06\n", 692 | "2.41676e+06\n", 693 | "2.10562e+06\n", 694 | "2.06967e+06\n", 695 | "2.26471e+06\n", 696 | "2.01519e+06\n", 697 | "2.21545e+06\n", 698 | "2.00595e+06\n", 699 | "1.9697e+06\n", 700 | "2.05301e+06\n", 701 | "1.97384e+06\n", 702 | "2.14254e+06\n", 703 | "2.20974e+06\n", 704 | "2.01116e+06\n", 705 | "2.16271e+06\n", 706 | "2.15502e+06\n", 707 | "1.83303e+06\n", 708 | "2.21899e+06\n", 709 | "1.84089e+06\n", 710 | "2.0579e+06\n", 711 | "1.96048e+06\n", 712 | "2.05482e+06\n", 713 | "2.00343e+06\n", 714 | "1.92695e+06\n", 715 | "2.00567e+06\n", 716 | "1.91174e+06\n", 717 | "2.1585e+06\n", 718 | "1.77096e+06\n", 719 | "1.95128e+06\n", 720 | "1.98419e+06\n", 721 | "2.01649e+06\n", 722 | "2.01482e+06\n", 723 | "2.06095e+06\n", 724 | "1.94039e+06\n", 725 | "2.09409e+06\n", 726 | "1.83339e+06\n", 727 | "2.316e+06\n", 728 | "2.02485e+06\n", 729 | "1.93106e+06\n", 730 | "2.21034e+06\n", 731 | "2.18385e+06\n", 732 | "2.17211e+06\n", 733 | "2.00824e+06\n", 734 | "2.05121e+06\n", 735 | "1.92578e+06\n", 736 | "1.89011e+06\n", 737 | "1.96128e+06\n", 738 | "2.10993e+06\n", 739 | "1.95284e+06\n", 740 | "1.97745e+06\n", 741 | "1.91478e+06\n", 742 | "1.89365e+06\n", 743 | "2.16375e+06\n", 744 | "2.12979e+06\n", 745 | "2.03724e+06\n", 746 | "2.09712e+06\n", 747 | "1.72613e+06\n", 748 | "2.10846e+06\n", 749 | "2.11904e+06\n", 750 | "1.948e+06\n", 751 | "2.02451e+06\n", 752 | "1.81043e+06\n", 753 | "1.86072e+06\n", 754 | "2.02043e+06\n", 755 | "2.16489e+06\n", 756 | "1.9269e+06\n", 757 | "2.21788e+06\n", 758 | "1.96544e+06\n", 759 | "1.9333e+06\n", 760 | "2.08736e+06\n", 761 | "2.01953e+06\n", 762 | "1.96375e+06\n", 763 | "2.07949e+06\n", 764 | "1.90033e+06\n", 765 | "1.85397e+06\n", 766 | "1.81195e+06\n", 767 | "2.03793e+06\n", 768 | "2.00546e+06\n", 769 | "2.06524e+06\n", 770 | "1.9061e+06\n", 771 | "2.01634e+06\n", 772 | "1.84958e+06\n", 773 | "1.91682e+06\n", 774 | "2.12577e+06\n", 775 | "1.94529e+06\n", 776 | "2.01208e+06\n", 777 | "1.85676e+06\n", 778 | "1.88109e+06\n", 779 | "1.8846e+06\n", 780 | "1.95878e+06\n", 781 | "1.86846e+06\n", 782 | "2.0202e+06\n", 783 | "1.94568e+06\n", 784 | "1.97567e+06\n", 785 | "1.95601e+06\n", 786 | "1.87193e+06\n", 787 | "2.03693e+06\n", 788 | "1.66641e+06\n", 789 | "1.82568e+06\n", 790 | "1.86975e+06\n", 791 | "2.06113e+06\n", 792 | "1.83814e+06\n", 793 | "1.80009e+06\n", 794 | "1.82125e+06\n", 795 | "1.84622e+06\n", 796 | "1.91394e+06\n", 797 | "2.03654e+06\n", 798 | "1.9048e+06\n", 799 | "2.08284e+06\n", 800 | "1.98599e+06\n", 801 | "1.78153e+06\n", 802 | "1.8436e+06\n", 803 | "1.89547e+06\n", 804 | "2.26996e+06\n", 805 | "1.92931e+06\n", 806 | "2.02713e+06\n", 807 | "2.19537e+06\n", 808 | "1.76284e+06\n", 809 | "1.99358e+06\n", 810 | "1.64292e+06\n", 811 | "1.7837e+06\n", 812 | "1.94135e+06\n", 813 | "1.68807e+06\n", 814 | "1.87227e+06\n", 815 | "2.02543e+06\n", 816 | "1.87374e+06\n", 817 | "2.07681e+06\n", 818 | "2.04221e+06\n", 819 | "2.05031e+06\n", 820 | "1.86743e+06\n", 821 | "1.93306e+06\n", 822 | "1.80954e+06\n", 823 | "1.69309e+06\n", 824 | "1.99876e+06\n", 825 | "1.73244e+06\n", 826 | "1.74135e+06\n", 827 | "1.91624e+06\n", 828 | "2.0926e+06\n", 829 | "1.80757e+06\n", 830 | "2.00978e+06\n", 831 | "1.72931e+06\n", 832 | "1.76419e+06\n", 833 | "1.90187e+06\n", 834 | "1.96696e+06\n", 835 | "1.83468e+06\n", 836 | "1.75355e+06\n", 837 | "1.6695e+06\n", 838 | "1.89424e+06\n", 839 | "1.8585e+06\n", 840 | "2.06588e+06\n", 841 | "1.81253e+06\n", 842 | "1.74599e+06\n", 843 | "1.84886e+06\n", 844 | "1.96784e+06\n", 845 | "1.89357e+06\n", 846 | "1.94487e+06\n", 847 | "1.75876e+06\n", 848 | "1.65248e+06\n", 849 | "1.77267e+06\n", 850 | "1.60278e+06\n", 851 | "1.79852e+06\n", 852 | "1.97353e+06\n", 853 | "1.71115e+06\n", 854 | "1.849e+06\n", 855 | "1.91462e+06\n", 856 | "1.88385e+06\n", 857 | "2.00161e+06\n", 858 | "1.79726e+06\n", 859 | "1.91972e+06\n", 860 | "1.78144e+06\n", 861 | "1.78448e+06\n", 862 | "1.91225e+06\n", 863 | "1.79337e+06\n", 864 | "1.85673e+06\n", 865 | "1.89331e+06\n", 866 | "1.72505e+06\n", 867 | "1.97252e+06\n", 868 | "1.81042e+06\n", 869 | "1.74078e+06\n", 870 | "1.6816e+06\n", 871 | "1.85483e+06\n", 872 | "1.65616e+06\n", 873 | "1.93177e+06\n", 874 | "1.77377e+06\n", 875 | "1.6367e+06\n", 876 | "1.81177e+06\n", 877 | "1.75343e+06\n", 878 | "1.8101e+06\n", 879 | "1.58939e+06\n", 880 | "1.62266e+06\n", 881 | "1.89291e+06\n", 882 | "1.83072e+06\n", 883 | "1.9542e+06\n", 884 | "1.72349e+06\n", 885 | "1.77785e+06\n", 886 | "1.75284e+06\n", 887 | "1.7304e+06\n", 888 | "1.91349e+06\n", 889 | "1.75855e+06\n", 890 | "1.58313e+06\n", 891 | "1.71256e+06\n", 892 | "1.86418e+06\n", 893 | "1.86264e+06\n", 894 | "1.89268e+06\n", 895 | "1.99144e+06\n", 896 | "1.73327e+06\n", 897 | "1.87112e+06\n", 898 | "1.67488e+06\n", 899 | "1.86644e+06\n", 900 | "1.91273e+06\n", 901 | "1.72277e+06\n", 902 | "1.73226e+06\n", 903 | "1.75132e+06\n", 904 | "1.81514e+06\n", 905 | "1.64353e+06\n", 906 | "1.60016e+06\n", 907 | "1.71197e+06\n", 908 | "1.77534e+06\n", 909 | "1.83148e+06\n", 910 | "1.93861e+06\n", 911 | "1.78132e+06\n", 912 | "1.75223e+06\n", 913 | "1.90546e+06\n", 914 | "1.72876e+06\n", 915 | "1.85251e+06\n", 916 | "1.5776e+06\n", 917 | "1.63816e+06\n", 918 | "1.77009e+06\n", 919 | "1.80358e+06\n", 920 | "1.60131e+06\n", 921 | "1.69186e+06\n", 922 | "1.6854e+06\n", 923 | "1.66458e+06\n", 924 | "1.80586e+06\n", 925 | "1.58416e+06\n", 926 | "1.73365e+06\n", 927 | "1.89232e+06\n", 928 | "1.59774e+06\n", 929 | "1.65748e+06\n", 930 | "1.62077e+06\n", 931 | "1.68111e+06\n", 932 | "1.71427e+06\n", 933 | "1.73181e+06\n", 934 | "1.86657e+06\n", 935 | "1.84007e+06\n", 936 | "1.71243e+06\n", 937 | "1.89197e+06\n", 938 | "1.75037e+06\n", 939 | "1.71707e+06\n", 940 | "1.65104e+06\n", 941 | "1.72606e+06\n", 942 | "1.6638e+06\n", 943 | "1.55797e+06\n", 944 | "1.69882e+06\n", 945 | "1.7289e+06\n", 946 | "1.4616e+06\n", 947 | "1.71243e+06\n", 948 | "1.87739e+06\n", 949 | "1.7327e+06\n", 950 | "1.82912e+06\n", 951 | "1.88278e+06\n", 952 | "1.67932e+06\n", 953 | "1.80275e+06\n", 954 | "1.75748e+06\n", 955 | "1.68215e+06\n", 956 | "1.68128e+06\n", 957 | "1.74823e+06\n", 958 | "1.64172e+06\n", 959 | "1.65909e+06\n", 960 | "1.85571e+06\n", 961 | "1.54681e+06\n", 962 | "1.77809e+06\n", 963 | "1.60919e+06\n", 964 | "1.73291e+06\n", 965 | "1.60452e+06\n", 966 | "1.75187e+06\n", 967 | "1.55642e+06\n", 968 | "1.64758e+06\n", 969 | "1.61989e+06\n", 970 | "1.73654e+06\n", 971 | "1.7267e+06\n", 972 | "1.52507e+06\n", 973 | "1.77411e+06\n", 974 | "1.5798e+06\n", 975 | "1.77472e+06\n", 976 | "1.67358e+06\n", 977 | "1.76103e+06\n", 978 | "1.59341e+06\n", 979 | "1.70558e+06\n", 980 | "1.76767e+06\n", 981 | "1.66647e+06\n", 982 | "1.58562e+06\n", 983 | "1.74169e+06\n", 984 | "1.68596e+06\n", 985 | "1.80671e+06\n", 986 | "1.63187e+06\n", 987 | "1.56945e+06\n", 988 | "1.68745e+06\n", 989 | "1.56728e+06\n", 990 | "1.59083e+06\n", 991 | "1.69317e+06\n", 992 | "1.75207e+06\n", 993 | "1.77906e+06\n", 994 | "1.7286e+06\n", 995 | "1.81511e+06\n", 996 | "1.61291e+06\n", 997 | "1.79015e+06\n", 998 | "1.80175e+06\n", 999 | "1.75647e+06\n", 1000 | "1.57833e+06\n", 1001 | "1.51464e+06\n", 1002 | "1.66931e+06\n", 1003 | "1.49117e+06\n", 1004 | "1.58909e+06\n", 1005 | "1.61494e+06\n", 1006 | "1.51524e+06\n", 1007 | "1.57864e+06\n", 1008 | "1.52599e+06\n", 1009 | "1.76529e+06\n", 1010 | "1.66485e+06\n", 1011 | "1.59218e+06\n", 1012 | "1.62364e+06\n", 1013 | "1.5007e+06\n", 1014 | "1.55731e+06\n", 1015 | "1.52107e+06\n", 1016 | "1.6285e+06\n", 1017 | "1.566e+06\n", 1018 | "1.73527e+06\n", 1019 | "1.45545e+06\n", 1020 | "1.55773e+06\n", 1021 | "1.56766e+06\n", 1022 | "1.58881e+06\n" 1023 | ] 1024 | } 1025 | ], 1026 | "source": [ 1027 | "with tf.Session() as sess:\n", 1028 | " sess.run(tf.initialize_all_variables())\n", 1029 | " for _ in range(500):\n", 1030 | " # select random mini-batch\n", 1031 | " indices = np.random.choice(n_samples, batch_size)\n", 1032 | " x_batch = x_data[indices]\n", 1033 | " y_batch = y_data[indices]\n", 1034 | " _, loss_val = sess.run([opt_operation, loss], feed_dict={x_input: x_batch, y_input: y_batch})\n", 1035 | " print(loss_val)" 1036 | ] 1037 | } 1038 | ], 1039 | "metadata": { 1040 | "kernelspec": { 1041 | "display_name": "Python 2", 1042 | "language": "python", 1043 | "name": "python2" 1044 | }, 1045 | "language_info": { 1046 | "codemirror_mode": { 1047 | "name": "ipython", 1048 | "version": 2 1049 | }, 1050 | "file_extension": ".py", 1051 | "mimetype": "text/x-python", 1052 | "name": "python", 1053 | "nbconvert_exporter": "python", 1054 | "pygments_lexer": "ipython2", 1055 | "version": "2.7.12" 1056 | } 1057 | }, 1058 | "nbformat": 4, 1059 | "nbformat_minor": 1 1060 | } 1061 | -------------------------------------------------------------------------------- /Task4-Tensorflow-Intro/Tensorflow-Softmax-Classifier-Toy-Data.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "### Imports" 8 | ] 9 | }, 10 | { 11 | "cell_type": "code", 12 | "execution_count": 1, 13 | "metadata": { 14 | "collapsed": true 15 | }, 16 | "outputs": [], 17 | "source": [ 18 | "import tensorflow as tf\n", 19 | "%matplotlib inline\n", 20 | "from matplotlib import pyplot as plt\n", 21 | "import numpy as np\n", 22 | "from sklearn.preprocessing import LabelBinarizer\n", 23 | "plt.rcParams['figure.figsize'] = (10.0, 8.0) # set default size of plots\n", 24 | "plt.rcParams['image.interpolation'] = 'nearest'\n", 25 | "plt.rcParams['image.cmap'] = 'gray'" 26 | ] 27 | }, 28 | { 29 | "cell_type": "code", 30 | "execution_count": 2, 31 | "metadata": { 32 | "collapsed": false 33 | }, 34 | "outputs": [ 35 | { 36 | "data": { 37 | "text/plain": [ 38 | "" 39 | ] 40 | }, 41 | "execution_count": 2, 42 | "metadata": {}, 43 | "output_type": "execute_result" 44 | } 45 | ], 46 | "source": [ 47 | "tf.InteractiveSession()" 48 | ] 49 | }, 50 | { 51 | "cell_type": "markdown", 52 | "metadata": {}, 53 | "source": [ 54 | "### Generating Data" 55 | ] 56 | }, 57 | { 58 | "cell_type": "code", 59 | "execution_count": 3, 60 | "metadata": { 61 | "collapsed": false 62 | }, 63 | "outputs": [], 64 | "source": [ 65 | "N = 100 # number of points per class\n", 66 | "D = 2 # dimensionality\n", 67 | "K = 3 # number of classes\n", 68 | "X = np.zeros((N*K,D), dtype='float32') # data matrix (each row = single example)\n", 69 | "y = np.zeros(N*K, dtype='int32') # class labels" 70 | ] 71 | }, 72 | { 73 | "cell_type": "code", 74 | "execution_count": 4, 75 | "metadata": { 76 | "collapsed": false 77 | }, 78 | "outputs": [], 79 | "source": [ 80 | "for j in xrange(K):\n", 81 | " indices = range(N*j,N*(j+1))\n", 82 | " r = np.linspace(0.0,1,N) # radius\n", 83 | " t = np.linspace(j*4,(j+1)*4,N) + np.random.randn(N)*0.2 # theta\n", 84 | " X[indices] = np.c_[r*np.sin(t), r*np.cos(t)]\n", 85 | " y[indices] = j" 86 | ] 87 | }, 88 | { 89 | "cell_type": "code", 90 | "execution_count": 5, 91 | "metadata": { 92 | "collapsed": false 93 | }, 94 | "outputs": [], 95 | "source": [ 96 | "ohe = LabelBinarizer()\n", 97 | "y_binary = ohe.fit_transform(y)" 98 | ] 99 | }, 100 | { 101 | "cell_type": "code", 102 | "execution_count": 6, 103 | "metadata": { 104 | "collapsed": false 105 | }, 106 | "outputs": [ 107 | { 108 | "data": { 109 | "text/plain": [ 110 | "" 111 | ] 112 | }, 113 | "execution_count": 6, 114 | "metadata": {}, 115 | "output_type": "execute_result" 116 | }, 117 | { 118 | "data": { 119 | "image/png": "iVBORw0KGgoAAAANSUhEUgAAA00AAAKaCAYAAADiT7IPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niEgO/j8uy8Ka4aXJPAAAAABJRU5ErkJggg==\n", 120 | "text/plain": [ 121 | "" 122 | ] 123 | }, 124 | "metadata": {}, 125 | "output_type": "display_data" 126 | } 127 | ], 128 | "source": [ 129 | "# lets visualize the data:\n", 130 | "plt.scatter(X[:, 0], X[:, 1], c=y, s=40, cmap=plt.cm.Spectral)" 131 | ] 132 | }, 133 | { 134 | "cell_type": "markdown", 135 | "metadata": {}, 136 | "source": [ 137 | "### Training a softmax classifier" 138 | ] 139 | }, 140 | { 141 | "cell_type": "markdown", 142 | "metadata": {}, 143 | "source": [ 144 | "#### Define placeholders for X and y inputs" 145 | ] 146 | }, 147 | { 148 | "cell_type": "code", 149 | "execution_count": 7, 150 | "metadata": { 151 | "collapsed": true 152 | }, 153 | "outputs": [], 154 | "source": [ 155 | "X_input = tf.placeholder(tf.float32, shape=X.shape)\n", 156 | "y_input = tf.placeholder(tf.float32, shape=y_binary.shape)" 157 | ] 158 | }, 159 | { 160 | "cell_type": "markdown", 161 | "metadata": {}, 162 | "source": [ 163 | "#### Initialize weights and bias" 164 | ] 165 | }, 166 | { 167 | "cell_type": "code", 168 | "execution_count": 8, 169 | "metadata": { 170 | "collapsed": false 171 | }, 172 | "outputs": [], 173 | "source": [ 174 | "# initialize weights\n", 175 | "W = tf.get_variable(\"weights\", (D,K), initializer=tf.random_normal_initializer())\n", 176 | "# initialize bias\n", 177 | "b = tf.get_variable(\"bias\", (1,K), initializer=tf.random_normal_initializer())" 178 | ] 179 | }, 180 | { 181 | "cell_type": "markdown", 182 | "metadata": {}, 183 | "source": [ 184 | "#### Compute loss" 185 | ] 186 | }, 187 | { 188 | "cell_type": "code", 189 | "execution_count": 9, 190 | "metadata": { 191 | "collapsed": false 192 | }, 193 | "outputs": [], 194 | "source": [ 195 | "# compute class scores (dim = (N*K,K))\n", 196 | "scores = tf.matmul(X_input, W) + b\n", 197 | "# normalize scores to probabilities (dim = (N*K,K))\n", 198 | "probabilities = tf.exp(scores)/tf.reduce_sum(tf.exp(scores), reduction_indices=1, keep_dims=True)\n", 199 | "# extract correct class probabilities (dim = (N*K,1))\n", 200 | "correct_logprobas = -(y_input * tf.log(probabilities))\n", 201 | "# compute the data loss\n", 202 | "data_loss = tf.reduce_sum(correct_logprobas)/(N*K)\n", 203 | "# compute the regularization loss\n", 204 | "reg_lambda = 1\n", 205 | "regularization_loss = 0.5*reg_lambda*tf.reduce_sum(W*W)\n", 206 | "loss = data_loss + regularization_loss" 207 | ] 208 | }, 209 | { 210 | "cell_type": "markdown", 211 | "metadata": {}, 212 | "source": [ 213 | "#### Define the loss optimizer" 214 | ] 215 | }, 216 | { 217 | "cell_type": "code", 218 | "execution_count": 10, 219 | "metadata": { 220 | "collapsed": false 221 | }, 222 | "outputs": [], 223 | "source": [ 224 | "opt = tf.train.GradientDescentOptimizer(1e-0)\n", 225 | "opt_operation = opt.minimize(loss)" 226 | ] 227 | }, 228 | { 229 | "cell_type": "markdown", 230 | "metadata": {}, 231 | "source": [ 232 | "#### Train and calculate accuracy" 233 | ] 234 | }, 235 | { 236 | "cell_type": "code", 237 | "execution_count": 11, 238 | "metadata": { 239 | "collapsed": false 240 | }, 241 | "outputs": [ 242 | { 243 | "name": "stdout", 244 | "output_type": "stream", 245 | "text": [ 246 | "training accuracy: 0.54\n", 247 | "----------Weights---------\n", 248 | "[array([[ 0.03963081, 0.0512507 , -0.09088154],\n", 249 | " [-0.08217616, 0.09520487, -0.01302874]], dtype=float32)]\n" 250 | ] 251 | } 252 | ], 253 | "source": [ 254 | "with tf.Session() as sess:\n", 255 | " sess.run(tf.initialize_all_variables())\n", 256 | " for _ in range(500):\n", 257 | " _, loss_val = sess.run([opt_operation, loss], feed_dict={X_input: X, y_input: y_binary}) \n", 258 | " scores = tf.matmul(X,W) + b\n", 259 | " predicted_class = tf.argmax(scores,1)\n", 260 | " print 'training accuracy: %.2f' % (np.mean(predicted_class.eval() == y))\n", 261 | " print(\"----------Weights---------\")\n", 262 | " print(sess.run([W]))" 263 | ] 264 | } 265 | ], 266 | "metadata": { 267 | "kernelspec": { 268 | "display_name": "Python 2", 269 | "language": "python", 270 | "name": "python2" 271 | }, 272 | "language_info": { 273 | "codemirror_mode": { 274 | "name": "ipython", 275 | "version": 2 276 | }, 277 | "file_extension": ".py", 278 | "mimetype": "text/x-python", 279 | "name": "python", 280 | "nbconvert_exporter": "python", 281 | "pygments_lexer": "ipython2", 282 | "version": "2.7.12" 283 | } 284 | }, 285 | "nbformat": 4, 286 | "nbformat_minor": 1 287 | } 288 | --------------------------------------------------------------------------------