├── images ├── ddpg.png └── deep-q-learning-algorithm.png ├── .gitignore ├── README.md ├── 71 [DDPG] Deep Deterministic Policy Gradients.ipynb ├── 01 Pytorch Getteing Started.ipynb ├── 02 [MNIST] Simple Forward and Backward.ipynb └── 03 [CNN] CIFAR-10 Classification.ipynb /images/ddpg.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AndersonJo/pytorch-examples/master/images/ddpg.png -------------------------------------------------------------------------------- /images/deep-q-learning-algorithm.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/AndersonJo/pytorch-examples/master/images/deep-q-learning-algorithm.png -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | # Project 2 | .idea 3 | .ipynb_checkpoints 4 | 5 | # Data 6 | data 7 | raw 8 | processed 9 | hymenoptera_data 10 | 11 | # IPython 12 | .ipynb_checkpoints 13 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Pytorch Examples 2 | 3 | The repository contains Pytorch examples using Jupyter Notebook for simplicity. 4 | 5 | 6 | ## Prerequisites 7 | 8 | 1. Python 3.3+ 9 | 2. [Pytorch](https://github.com/pytorch/pytorch) 10 | 3. [Torchvision](https://github.com/pytorch/vision/tree/master/torchvision) 11 | 12 | ## Examples 13 | 14 | 1. [Getting Started with Pytorch](https://github.com/AndersonJo/pytorch-examples/blob/master/01%20Pytorch%20Getteing%20Started.ipynb) 15 | 16 | 2. [[DeepLearning] MNIST Classification with Deep Neural Network](https://github.com/AndersonJo/pytorch-examples/blob/master/02%20%5BMNIST%5D%20Simple%20Forward%20and%20Backward.ipynb) 17 | 18 | 3. [[CNN] CIFAR-10 Classification](https://github.com/AndersonJo/pytorch-examples/blob/master/03%20%5BCNN%5D%20CIFAR-10%20Classification.ipynb) 19 | 20 | 4. [[GAN] Heuristic Method](https://github.com/AndersonJo/pytorch-examples/blob/master/50%20%5BHeuristic%20GAN%5D%20non-saturating%20game%20for%20MNIST.ipynb) 21 | -------------------------------------------------------------------------------- /71 [DDPG] Deep Deterministic Policy Gradients.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# Continuous Control with Deep Reinforcement Learning\n", 8 | "\n", 9 | "## DQN의 문제점 그리고 DDPG\n", 10 | "\n", 11 | "DQN은 많은 복잡한 문제들을 풀었지만, 오직 discrete 하며 low-dimensional action spaces의 문제를 갖은 문제를 해결할 수 있습니다.
\n", 12 | "즉 '왼쪽', '오른쪽', '점프', '앉기' 등등 서로 구분(discrete)가 명확하며, actions의 갯수 (action spaces)가 적은 문제를 말합니다.\n", 13 | "\n", 14 | "하지만 현실에서는 \"물리적 컨트롤\"을 요구하는 문제들이 많습니다.
\n", 15 | "손의 악력의 정도(continuous 값), 3차원의 방향에서 어느 방향으로 움직일지(너무나 많은 action spaces) 등등..
\n", 16 | "이런 문제에 DQN이 사용되기 쉽지 않습니다. 이유는 궁극적으로 DQN은 action-value function의 값을 극대화하는 action을 찾는 것이며,
\n", 17 | "이를 찾기 위해서는 모든 step마다 iterative optimization process를 거쳐야 하기 때문입니다. \n", 18 | "\n", 19 | "이를 해결하기 위한 한가지 방법은 **discretize the action space** 해주는 방법입니다.
\n", 20 | "예를 들어서 사람의 팔은 7도 정도 움직일수 있는데, 이것자체를 action으로 두는 것입니다.
\n", 21 | "각각의 각도마다 각각의 action을 만드는 방법인데.. 세밀하게 움직이게 하려면 할 수록 action space는 기하급수적으로 늘어나게 됩니다.\n", 22 | "\n", 23 | "특히 로보틱스 분야에서는 continuous action을 다루게 되는데, Google Deep Mind팀은 이를 해결하기 위해서 **Deep Deterministic Policy Gradients (DDPG)**라는 알고리즘을 만들었고, 해당 알고리즘은 **off-policy** 이며 **model-free** 라는 특징을 갖고 있습니다.
\n", 24 | "DDPG는 3개의 techniques를 합쳐서 만들어졌습니다." 25 | ] 26 | }, 27 | { 28 | "cell_type": "markdown", 29 | "metadata": {}, 30 | "source": [ 31 | "## Policy Network\n", 32 | "\n" 33 | ] 34 | }, 35 | { 36 | "cell_type": "markdown", 37 | "metadata": {}, 38 | "source": [ 39 | "### DDPG Algorithm\n", 40 | "\n", 41 | "Deep Deterministic policy gradient를 뜻하며 target network의 알고리즘을 포함하고 있습니다.\n", 42 | "\n", 43 | "\n", 44 | "\n" 45 | ] 46 | }, 47 | { 48 | "cell_type": "markdown", 49 | "metadata": {}, 50 | "source": [ 51 | "# References \n", 52 | "\n", 53 | "* [Target Network에 대한 자세한 Paper - CONTINUOUS CONTROL WITH DEEP REINFORCEMENT LEARNING](https://arxiv.org/pdf/1509.02971.pdf)\n", 54 | "* [Deep Deterministic Policy Gradients in TensorFlow](http://pemami4911.github.io/blog/2016/08/21/ddpg-rl.html)\n", 55 | "* [Using Keras and Deep Deterministic Policy Gradient to play TORCS](https://yanpanlau.github.io/2016/10/11/Torcs-Keras.html)" 56 | ] 57 | } 58 | ], 59 | "metadata": { 60 | "kernelspec": { 61 | "display_name": "Python 3", 62 | "language": "python", 63 | "name": "python3" 64 | }, 65 | "language_info": { 66 | "codemirror_mode": { 67 | "name": "ipython", 68 | "version": 3 69 | }, 70 | "file_extension": ".py", 71 | "mimetype": "text/x-python", 72 | "name": "python", 73 | "nbconvert_exporter": "python", 74 | "pygments_lexer": "ipython3", 75 | "version": "3.6.1" 76 | } 77 | }, 78 | "nbformat": 4, 79 | "nbformat_minor": 2 80 | } 81 | -------------------------------------------------------------------------------- /01 Pytorch Getteing Started.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 1, 6 | "metadata": { 7 | "collapsed": true, 8 | "deletable": true, 9 | "editable": true 10 | }, 11 | "outputs": [], 12 | "source": [ 13 | "from torch.autograd import Variable\n", 14 | "\n", 15 | "import torch\n", 16 | "import torch.nn as nn\n", 17 | "import torch.nn.functional as F\n", 18 | "import torch.optim as optim" 19 | ] 20 | }, 21 | { 22 | "cell_type": "markdown", 23 | "metadata": { 24 | "deletable": true, 25 | "editable": true 26 | }, 27 | "source": [ 28 | "### Neural Network Model\n", 29 | "\n", 30 | "nn.Module 을 사용하여 Neural Network 모델을 만들고, 반드시 forward(input)함수를 만들어줘야 합니다.
\n", 31 | "backward함수는 자동으로 만들어집니다.\n", 32 | "\n", 33 | "> torch.nn 전체 모듈 통들어서.. 오직 minibatch만 제공됩니다.
\n", 34 | "> 즉 single sample은 전혀 지원되지 않습니다.
\n", 35 | "> 예를 들어서 4D Tensor는 다음과 같이 사용되어야 합니다.

\n", 36 | "> nSamples x nChannels x Height x Width

\n", 37 | "> 만약 single sample을 사용하려고 한다면 **input.unsqueeze(0)** 를 사용해서 fake batch를 만들어 줍니다." 38 | ] 39 | }, 40 | { 41 | "cell_type": "code", 42 | "execution_count": 6, 43 | "metadata": { 44 | "collapsed": false, 45 | "deletable": true, 46 | "editable": true 47 | }, 48 | "outputs": [ 49 | { 50 | "name": "stdout", 51 | "output_type": "stream", 52 | "text": [ 53 | "Net (\n", 54 | " (conv1): Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1))\n", 55 | " (conv2): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))\n", 56 | " (fc1): Linear (400 -> 120)\n", 57 | " (fc2): Linear (120 -> 84)\n", 58 | " (fc3): Linear (84 -> 10)\n", 59 | ")\n" 60 | ] 61 | } 62 | ], 63 | "source": [ 64 | "class Net(nn.Module):\n", 65 | "\n", 66 | " def __init__(self):\n", 67 | " super(Net, self).__init__()\n", 68 | " # 1 input image channel, 6 output channels, 5x5 square convolution kernel\n", 69 | " self.conv1 = nn.Conv2d(1, 6, 5).cuda()\n", 70 | " self.conv2 = nn.Conv2d(6, 16, 5).cuda()\n", 71 | " self.fc1 = nn.Linear(16 * 5 * 5, 120).cuda() # an affine operation: y = Wx + b\n", 72 | " self.fc2 = nn.Linear(120, 84).cuda()\n", 73 | " self.fc3 = nn.Linear(84, 10).cuda()\n", 74 | "\n", 75 | " def forward(self, x):\n", 76 | " x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2)) # Max pooling over a (2, 2) window\n", 77 | " x = F.max_pool2d(F.relu(self.conv2(x)), 2) # If the size is a square you can only specify a single number\n", 78 | " x = x.view(-1, self.num_flat_features(x))\n", 79 | " x = F.relu(self.fc1(x))\n", 80 | " x = F.relu(self.fc2(x))\n", 81 | " x = self.fc3(x)\n", 82 | " return x\n", 83 | "\n", 84 | " def num_flat_features(self, x):\n", 85 | " size = x.size()[1:] # all dimensions except the batch dimension\n", 86 | " num_features = 1\n", 87 | " for s in size:\n", 88 | " num_features *= s\n", 89 | " return num_features\n", 90 | "\n", 91 | "net = Net()\n", 92 | "print(net)" 93 | ] 94 | }, 95 | { 96 | "cell_type": "markdown", 97 | "metadata": { 98 | "deletable": true, 99 | "editable": true 100 | }, 101 | "source": [ 102 | "### Predict\n", 103 | "\n", 104 | "GPU에서 처리한 값을 CPU로 불러와서 처리해야 하기 때문에 .cpu 함수를 사용합니다." 105 | ] 106 | }, 107 | { 108 | "cell_type": "code", 109 | "execution_count": 161, 110 | "metadata": { 111 | "collapsed": false, 112 | "deletable": true, 113 | "editable": true 114 | }, 115 | "outputs": [ 116 | { 117 | "data": { 118 | "text/plain": [ 119 | "array([[-0.04526739, 0.01011707, 0.08764283, -0.07252117, 0.08807731,\n", 120 | " 0.00853545, -0.01395075, 0.0785292 , -0.06747445, 0.0142101 ]], dtype=float32)" 121 | ] 122 | }, 123 | "execution_count": 161, 124 | "metadata": {}, 125 | "output_type": "execute_result" 126 | } 127 | ], 128 | "source": [ 129 | "input = Variable(torch.randn(1, 1, 32, 32)).cuda()\n", 130 | "output = net(input)\n", 131 | "\n", 132 | "output.data.cpu().numpy()" 133 | ] 134 | }, 135 | { 136 | "cell_type": "markdown", 137 | "metadata": { 138 | "deletable": true, 139 | "editable": true 140 | }, 141 | "source": [ 142 | "### Loss function" 143 | ] 144 | }, 145 | { 146 | "cell_type": "code", 147 | "execution_count": 162, 148 | "metadata": { 149 | "collapsed": false, 150 | "deletable": true, 151 | "editable": true 152 | }, 153 | "outputs": [ 154 | { 155 | "name": "stdout", 156 | "output_type": "stream", 157 | "text": [ 158 | "Variable containing:\n", 159 | " 28.4200\n", 160 | "[torch.cuda.FloatTensor of size 1 (GPU 0)]\n", 161 | "\n" 162 | ] 163 | } 164 | ], 165 | "source": [ 166 | "target = Variable(torch.arange(0, 10)).cuda()\n", 167 | "criterion = nn.MSELoss()\n", 168 | "loss = criterion(output, target)\n", 169 | "print(loss)" 170 | ] 171 | }, 172 | { 173 | "cell_type": "code", 174 | "execution_count": 163, 175 | "metadata": { 176 | "collapsed": false, 177 | "deletable": true, 178 | "editable": true 179 | }, 180 | "outputs": [ 181 | { 182 | "name": "stdout", 183 | "output_type": "stream", 184 | "text": [ 185 | "\n", 186 | "\n", 187 | "\n" 188 | ] 189 | } 190 | ], 191 | "source": [ 192 | "print(loss.creator) # MSELoss\n", 193 | "print(loss.creator.previous_functions[0][0]) # Linear\n", 194 | "print(loss.creator.previous_functions[0][0].previous_functions[0][0]) # ReLU" 195 | ] 196 | }, 197 | { 198 | "cell_type": "markdown", 199 | "metadata": { 200 | "deletable": true, 201 | "editable": true 202 | }, 203 | "source": [ 204 | "### Backpropagation\n", 205 | "\n", 206 | "backpropagate the error 하기전에 반드시 .zero_grad()를 사용해서 gradients값을 0으로 만들어줘야 합니다.
\n", 207 | "이렇게 하지 않고, .backward()를 사용시 gradients값들이 accumulate되게 됩니다." 208 | ] 209 | }, 210 | { 211 | "cell_type": "code", 212 | "execution_count": 164, 213 | "metadata": { 214 | "collapsed": false, 215 | "deletable": true, 216 | "editable": true 217 | }, 218 | "outputs": [ 219 | { 220 | "name": "stdout", 221 | "output_type": "stream", 222 | "text": [ 223 | "conv1.bias.grad before backward\n", 224 | "Variable containing:\n", 225 | " 0\n", 226 | " 0\n", 227 | " 0\n", 228 | " 0\n", 229 | " 0\n", 230 | " 0\n", 231 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]\n", 232 | "\n", 233 | "conv1.bias.grad after backward\n", 234 | "Variable containing:\n", 235 | "1.00000e-02 *\n", 236 | " 0.0102\n", 237 | " -5.8490\n", 238 | " 4.7722\n", 239 | " 1.1550\n", 240 | " -2.7535\n", 241 | " 0.2246\n", 242 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]\n", 243 | "\n" 244 | ] 245 | } 246 | ], 247 | "source": [ 248 | "net.zero_grad() \n", 249 | "\n", 250 | "print('conv1.bias.grad before backward')\n", 251 | "print(net.conv1.bias.grad)\n", 252 | "\n", 253 | "loss.backward()\n", 254 | "\n", 255 | "print('conv1.bias.grad after backward')\n", 256 | "print(net.conv1.bias.grad)" 257 | ] 258 | }, 259 | { 260 | "cell_type": "markdown", 261 | "metadata": { 262 | "deletable": true, 263 | "editable": true 264 | }, 265 | "source": [ 266 | "### Update the weights\n", 267 | "\n", 268 | "가장 간단한 Stochastic Gradient Descent에 따르면 update는 다음과 같습니다.\n", 269 | "\n", 270 | "$$ \\text{weight} = \\text{weight} - \\text{learning rate} * \\text{gradient} $$\n", 271 | "\n", 272 | "파이썬에서 간단하게 하려면 다음과 같이 합니다." 273 | ] 274 | }, 275 | { 276 | "cell_type": "code", 277 | "execution_count": 165, 278 | "metadata": { 279 | "collapsed": true, 280 | "deletable": true, 281 | "editable": true 282 | }, 283 | "outputs": [], 284 | "source": [ 285 | "learning_rate = 0.01\n", 286 | "for f in net.parameters():\n", 287 | " f.data.sub_(f.grad.data * learning_rate)" 288 | ] 289 | }, 290 | { 291 | "cell_type": "markdown", 292 | "metadata": { 293 | "deletable": true, 294 | "editable": true 295 | }, 296 | "source": [ 297 | "하지만 SGD, Nesterov-SGD, Adam, RMSProp 같은 다양한 방법으로의 update가 존재합니다.
\n", 298 | "torch.optim 모듈은 이미 다양한 optimizers들을 제공하고 있습니다." 299 | ] 300 | }, 301 | { 302 | "cell_type": "code", 303 | "execution_count": 167, 304 | "metadata": { 305 | "collapsed": true, 306 | "deletable": true, 307 | "editable": true 308 | }, 309 | "outputs": [], 310 | "source": [ 311 | "optimizer = optim.SGD(net.parameters(), lr=0.01)\n", 312 | "\n", 313 | "# In your training loop\n", 314 | "# loop 안에서.. \n", 315 | "optimizer.zero_grad()\n", 316 | "output = net(input)\n", 317 | "loss = criterion(output, target)\n", 318 | "loss.backward()\n", 319 | "optimizer.step() # update를 함" 320 | ] 321 | }, 322 | { 323 | "cell_type": "markdown", 324 | "metadata": { 325 | "deletable": true, 326 | "editable": true 327 | }, 328 | "source": [ 329 | "### Learnable Parameters\n", 330 | "\n", 331 | "Learnable Parameters는 parameters() 함수를 통해서 얻을 수 있습니다." 332 | ] 333 | }, 334 | { 335 | "cell_type": "code", 336 | "execution_count": 7, 337 | "metadata": { 338 | "collapsed": false, 339 | "deletable": true, 340 | "editable": true 341 | }, 342 | "outputs": [ 343 | { 344 | "name": "stdout", 345 | "output_type": "stream", 346 | "text": [ 347 | "torch.Size([6, 1, 5, 5])\n", 348 | "torch.Size([6])\n", 349 | "torch.Size([16, 6, 5, 5])\n", 350 | "torch.Size([16])\n", 351 | "torch.Size([120, 400])\n", 352 | "torch.Size([120])\n", 353 | "torch.Size([84, 120])\n", 354 | "torch.Size([84])\n", 355 | "torch.Size([10, 84])\n", 356 | "torch.Size([10])\n", 357 | "\n", 358 | "Parameter containing:\n", 359 | " 3.6837e-02 -2.7158e-02 -2.4049e-02 ... 2.8363e-02 2.7047e-02 -2.0186e-02\n", 360 | " 2.6292e-02 2.1365e-04 4.3966e-02 ... 1.8812e-02 -3.9648e-02 -3.7087e-02\n", 361 | "-3.0545e-02 -2.0370e-02 -4.4853e-03 ... -1.6517e-02 -4.8652e-02 -6.4337e-03\n", 362 | " ... ⋱ ... \n", 363 | "-2.9419e-02 -2.1754e-03 -1.3552e-02 ... -1.5166e-02 2.2346e-03 3.7931e-02\n", 364 | " 3.8603e-02 -3.4742e-03 2.1454e-02 ... -1.6677e-02 -4.9337e-03 9.5037e-03\n", 365 | " 1.9295e-02 4.7208e-02 1.8468e-02 ... -1.0812e-02 5.0891e-03 2.5843e-02\n", 366 | "[torch.cuda.FloatTensor of size 120x400 (GPU 0)]\n", 367 | "\n" 368 | ] 369 | } 370 | ], 371 | "source": [ 372 | "params = list(net.parameters())\n", 373 | "for param in params:\n", 374 | " print(param.size())\n", 375 | "\n", 376 | "print()\n", 377 | "print(params[4])" 378 | ] 379 | }, 380 | { 381 | "cell_type": "markdown", 382 | "metadata": { 383 | "deletable": true, 384 | "editable": true 385 | }, 386 | "source": [ 387 | "### Gradients값 random으로 만들기" 388 | ] 389 | }, 390 | { 391 | "cell_type": "code", 392 | "execution_count": 113, 393 | "metadata": { 394 | "collapsed": false, 395 | "deletable": true, 396 | "editable": true 397 | }, 398 | "outputs": [ 399 | { 400 | "name": "stdout", 401 | "output_type": "stream", 402 | "text": [ 403 | "grad Variable containing:\n", 404 | "-0.0530\n", 405 | "-0.0365\n", 406 | "-0.0691\n", 407 | "-0.1326\n", 408 | " 0.0075\n", 409 | " 0.0064\n", 410 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]\n", 411 | "\n", 412 | "data \n", 413 | "-0.0331\n", 414 | "-0.1378\n", 415 | " 0.1596\n", 416 | "-0.0864\n", 417 | " 0.0523\n", 418 | " 0.1236\n", 419 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]\n", 420 | "\n" 421 | ] 422 | } 423 | ], 424 | "source": [ 425 | "output.backward(torch.randn(1, 10).cuda())\n", 426 | "params = list(net.parameters())\n", 427 | "print('grad', params[1].grad)\n", 428 | "print('data', params[1].data)" 429 | ] 430 | }, 431 | { 432 | "cell_type": "markdown", 433 | "metadata": { 434 | "deletable": true, 435 | "editable": true 436 | }, 437 | "source": [ 438 | "### Gradients값 zero로 만들기\n", 439 | "\n", 440 | ".zero_grad를 사용하게 되면 gradients값을 zero로 만듭니다." 441 | ] 442 | }, 443 | { 444 | "cell_type": "code", 445 | "execution_count": 114, 446 | "metadata": { 447 | "collapsed": false, 448 | "deletable": true, 449 | "editable": true 450 | }, 451 | "outputs": [ 452 | { 453 | "name": "stdout", 454 | "output_type": "stream", 455 | "text": [ 456 | "grad Variable containing:\n", 457 | " 0\n", 458 | " 0\n", 459 | " 0\n", 460 | " 0\n", 461 | " 0\n", 462 | " 0\n", 463 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]\n", 464 | "\n", 465 | "data \n", 466 | "-0.0331\n", 467 | "-0.1378\n", 468 | " 0.1596\n", 469 | "-0.0864\n", 470 | " 0.0523\n", 471 | " 0.1236\n", 472 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]\n", 473 | "\n" 474 | ] 475 | } 476 | ], 477 | "source": [ 478 | "net.zero_grad()\n", 479 | "params = list(net.parameters())\n", 480 | "print('grad', params[1].grad)\n", 481 | "print('data', params[1].data)" 482 | ] 483 | }, 484 | { 485 | "cell_type": "code", 486 | "execution_count": 115, 487 | "metadata": { 488 | "collapsed": false, 489 | "deletable": true, 490 | "editable": true 491 | }, 492 | "outputs": [ 493 | { 494 | "data": { 495 | "text/plain": [ 496 | "\n", 497 | "-0.0331\n", 498 | "-0.1378\n", 499 | " 0.1596\n", 500 | "-0.0864\n", 501 | " 0.0523\n", 502 | " 0.1236\n", 503 | "[torch.cuda.FloatTensor of size 6 (GPU 0)]" 504 | ] 505 | }, 506 | "execution_count": 115, 507 | "metadata": {}, 508 | "output_type": "execute_result" 509 | } 510 | ], 511 | "source": [ 512 | "a = params[1]\n", 513 | "a.data" 514 | ] 515 | } 516 | ], 517 | "metadata": { 518 | "kernelspec": { 519 | "display_name": "Python 3", 520 | "language": "python", 521 | "name": "python3" 522 | }, 523 | "language_info": { 524 | "codemirror_mode": { 525 | "name": "ipython", 526 | "version": 3 527 | }, 528 | "file_extension": ".py", 529 | "mimetype": "text/x-python", 530 | "name": "python", 531 | "nbconvert_exporter": "python", 532 | "pygments_lexer": "ipython3", 533 | "version": "3.6.0" 534 | } 535 | }, 536 | "nbformat": 4, 537 | "nbformat_minor": 2 538 | } 539 | -------------------------------------------------------------------------------- /02 [MNIST] Simple Forward and Backward.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 1, 6 | "metadata": { 7 | "collapsed": false, 8 | "deletable": true, 9 | "editable": true 10 | }, 11 | "outputs": [ 12 | { 13 | "name": "stdout", 14 | "output_type": "stream", 15 | "text": [ 16 | "Populating the interactive namespace from numpy and matplotlib\n" 17 | ] 18 | } 19 | ], 20 | "source": [ 21 | "%pylab inline\n", 22 | "import numpy as np\n", 23 | "import torch\n", 24 | "import torch.nn as nn\n", 25 | "import torch.nn.functional as F\n", 26 | "import torch.utils.data.dataloader as dataloader\n", 27 | "import torch.optim as optim\n", 28 | "\n", 29 | "from torch.utils.data import TensorDataset\n", 30 | "from torch.autograd import Variable\n", 31 | "from torchvision import transforms\n", 32 | "from torchvision.datasets import MNIST" 33 | ] 34 | }, 35 | { 36 | "cell_type": "markdown", 37 | "metadata": { 38 | "deletable": true, 39 | "editable": true 40 | }, 41 | "source": [ 42 | "## Data\n", 43 | "\n", 44 | "* [Pytorch Transform Documentation](http://pytorch.org/docs/torchvision/transforms.html)\n", 45 | "\n", 46 | "\n", 47 | "1. **torchvision.transforms.Compose:** 여러개의 tranforms을 실행합니다. \n", 48 | "2. **torchvision.transforms.ToTensor:** PIL.Image 또는 [0, 255] range의 Numpy array(H x W x C)를 (C x H x W)의 **[0.0, 1.0] range**를 갖은 torch.FloatTensor로 변형시킵니다.
여기서 포인트가 0에서 1사이의 값을 갖은 값으로 normalization이 포함되있습니다. \n", 49 | "3. **dataloader.DataLoader:** 사용하여 training시킬때 1개의 batch를 가져올때 shape이 **torch.Size([64, 1, 28, 28])** 이렇게 나옵니다. " 50 | ] 51 | }, 52 | { 53 | "cell_type": "code", 54 | "execution_count": 2, 55 | "metadata": { 56 | "collapsed": false, 57 | "deletable": true, 58 | "editable": true 59 | }, 60 | "outputs": [], 61 | "source": [ 62 | "train = MNIST('./data', train=True, download=True, transform=transforms.Compose([\n", 63 | " transforms.ToTensor(), # ToTensor does min-max normalization. \n", 64 | "]), )\n", 65 | "\n", 66 | "test = MNIST('./data', train=False, download=True, transform=transforms.Compose([\n", 67 | " transforms.ToTensor(), # ToTensor does min-max normalization. \n", 68 | "]), )\n", 69 | "\n", 70 | "# Create DataLoader\n", 71 | "dataloader_args = dict(shuffle=True, batch_size=64,num_workers=1, pin_memory=True)\n", 72 | "train_loader = dataloader.DataLoader(train, **dataloader_args)\n", 73 | "test_loader = dataloader.DataLoader(test, **dataloader_args)" 74 | ] 75 | }, 76 | { 77 | "cell_type": "code", 78 | "execution_count": 3, 79 | "metadata": { 80 | "collapsed": false, 81 | "deletable": true, 82 | "editable": true 83 | }, 84 | "outputs": [ 85 | { 86 | "name": "stdout", 87 | "output_type": "stream", 88 | "text": [ 89 | "[Train]\n", 90 | " - Numpy Shape: (60000, 28, 28)\n", 91 | " - Tensor Shape: torch.Size([60000, 28, 28])\n", 92 | " - Transformed Shape: torch.Size([28, 60000, 28])\n", 93 | " - min: 0.0\n", 94 | " - max: 1.0\n", 95 | " - mean: 0.13066047740240005\n", 96 | " - std: 0.3081078089011192\n", 97 | " - var: 0.0949304219058486\n" 98 | ] 99 | } 100 | ], 101 | "source": [ 102 | "train_data = train.train_data\n", 103 | "train_data = train.transform(train_data.numpy())\n", 104 | "\n", 105 | "print('[Train]')\n", 106 | "print(' - Numpy Shape:', train.train_data.cpu().numpy().shape)\n", 107 | "print(' - Tensor Shape:', train.train_data.size())\n", 108 | "print(' - Transformed Shape:', train_data.size())\n", 109 | "print(' - min:', torch.min(train_data))\n", 110 | "print(' - max:', torch.max(train_data))\n", 111 | "print(' - mean:', torch.mean(train_data))\n", 112 | "print(' - std:', torch.std(train_data))\n", 113 | "print(' - var:', torch.var(train_data))" 114 | ] 115 | }, 116 | { 117 | "cell_type": "markdown", 118 | "metadata": { 119 | "deletable": true, 120 | "editable": true 121 | }, 122 | "source": [ 123 | "## Model" 124 | ] 125 | }, 126 | { 127 | "cell_type": "code", 128 | "execution_count": 4, 129 | "metadata": { 130 | "collapsed": false, 131 | "deletable": true, 132 | "editable": true 133 | }, 134 | "outputs": [], 135 | "source": [ 136 | "class Model(nn.Module):\n", 137 | " def __init__(self):\n", 138 | " super(Model, self).__init__()\n", 139 | " \n", 140 | " self.fc1 = nn.Linear(784, 548)\n", 141 | " self.bc1 = nn.BatchNorm1d(548)\n", 142 | " \n", 143 | " self.fc2 = nn.Linear(548, 252)\n", 144 | " self.bc2 = nn.BatchNorm1d(252)\n", 145 | " \n", 146 | " self.fc3 = nn.Linear(252, 10)\n", 147 | " \n", 148 | " \n", 149 | " def forward(self, x):\n", 150 | " x = x.view((-1, 784))\n", 151 | " h = self.fc1(x)\n", 152 | " h = self.bc1(h)\n", 153 | " h = F.relu(h)\n", 154 | " h = F.dropout(h, p=0.5, training=self.training)\n", 155 | " \n", 156 | " h = self.fc2(h)\n", 157 | " h = self.bc2(h)\n", 158 | " h = F.relu(h)\n", 159 | " h = F.dropout(h, p=0.2, training=self.training)\n", 160 | " \n", 161 | " h = self.fc3(h)\n", 162 | " out = F.log_softmax(h)\n", 163 | " return out\n", 164 | "\n", 165 | "model = Model()\n", 166 | "model.cuda() # CUDA!\n", 167 | "optimizer = optim.Adam(model.parameters(), lr=0.001)" 168 | ] 169 | }, 170 | { 171 | "cell_type": "markdown", 172 | "metadata": { 173 | "deletable": true, 174 | "editable": true 175 | }, 176 | "source": [ 177 | "## Train" 178 | ] 179 | }, 180 | { 181 | "cell_type": "code", 182 | "execution_count": 5, 183 | "metadata": { 184 | "collapsed": false, 185 | "deletable": true, 186 | "editable": true 187 | }, 188 | "outputs": [ 189 | { 190 | "name": "stdout", 191 | "output_type": "stream", 192 | "text": [ 193 | " Train Epoch: 0 [57664/60000 (96%)]\tLoss: 0.221251\n", 194 | " Train Epoch: 1 [57664/60000 (96%)]\tLoss: 0.163002\n", 195 | " Train Epoch: 2 [57664/60000 (96%)]\tLoss: 0.219655\n", 196 | " Train Epoch: 3 [57664/60000 (96%)]\tLoss: 0.141791\n", 197 | " Train Epoch: 4 [57664/60000 (96%)]\tLoss: 0.080657\n", 198 | " Train Epoch: 5 [57664/60000 (96%)]\tLoss: 0.053320\n", 199 | " Train Epoch: 6 [57664/60000 (96%)]\tLoss: 0.053066\n", 200 | " Train Epoch: 7 [57664/60000 (96%)]\tLoss: 0.035491\n", 201 | " Train Epoch: 8 [57664/60000 (96%)]\tLoss: 0.010800\n", 202 | " Train Epoch: 9 [57664/60000 (96%)]\tLoss: 0.099638\n", 203 | " Train Epoch: 10 [57664/60000 (96%)]\tLoss: 0.150071\n", 204 | " Train Epoch: 11 [57664/60000 (96%)]\tLoss: 0.010240\n", 205 | " Train Epoch: 12 [57664/60000 (96%)]\tLoss: 0.071190\n", 206 | " Train Epoch: 13 [57664/60000 (96%)]\tLoss: 0.006421\n", 207 | " Train Epoch: 14 [57664/60000 (96%)]\tLoss: 0.031608\n" 208 | ] 209 | } 210 | ], 211 | "source": [ 212 | "model.train()\n", 213 | "\n", 214 | "losses = []\n", 215 | "for epoch in range(15):\n", 216 | " for batch_idx, (data, target) in enumerate(train_loader):\n", 217 | " # Get Samples\n", 218 | " data, target = Variable(data.cuda()), Variable(target.cuda())\n", 219 | " \n", 220 | " # Init\n", 221 | " optimizer.zero_grad()\n", 222 | "\n", 223 | " # Predict\n", 224 | " y_pred = model(data) \n", 225 | "\n", 226 | " # Calculate loss\n", 227 | " loss = F.cross_entropy(y_pred, target)\n", 228 | " losses.append(loss.data[0])\n", 229 | " # Backpropagation\n", 230 | " loss.backward()\n", 231 | " optimizer.step()\n", 232 | " \n", 233 | " \n", 234 | " # Display\n", 235 | " if batch_idx % 100 == 1:\n", 236 | " print('\\r Train Epoch: {} [{}/{} ({:.0f}%)]\\tLoss: {:.6f}'.format(\n", 237 | " epoch, \n", 238 | " batch_idx * len(data), \n", 239 | " len(train_loader.dataset),\n", 240 | " 100. * batch_idx / len(train_loader), \n", 241 | " loss.data[0]), \n", 242 | " end='')\n", 243 | " \n", 244 | " print()" 245 | ] 246 | }, 247 | { 248 | "cell_type": "code", 249 | "execution_count": 6, 250 | "metadata": { 251 | "collapsed": false, 252 | "deletable": true, 253 | "editable": true 254 | }, 255 | "outputs": [ 256 | { 257 | "data": { 258 | "text/plain": [ 259 | "[]" 260 | ] 261 | }, 262 | "execution_count": 6, 263 | "metadata": {}, 264 | "output_type": "execute_result" 265 | }, 266 | { 267 | "data": { 268 | "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", 269 | "text/plain": [ 270 | "" 271 | ] 272 | }, 273 | "metadata": {}, 274 | "output_type": "display_data" 275 | } 276 | ], 277 | "source": [ 278 | "plot(losses)" 279 | ] 280 | }, 281 | { 282 | "cell_type": "markdown", 283 | "metadata": { 284 | "deletable": true, 285 | "editable": true 286 | }, 287 | "source": [ 288 | "## Evaluate" 289 | ] 290 | }, 291 | { 292 | "cell_type": "code", 293 | "execution_count": 7, 294 | "metadata": { 295 | "collapsed": false, 296 | "deletable": true, 297 | "editable": true 298 | }, 299 | "outputs": [ 300 | { 301 | "name": "stdout", 302 | "output_type": "stream", 303 | "text": [ 304 | "Accuracy: 0.9764\n" 305 | ] 306 | } 307 | ], 308 | "source": [ 309 | "evaluate_x = Variable(test_loader.dataset.test_data.type_as(torch.FloatTensor())).cuda()\n", 310 | "evaluate_y = Variable(test_loader.dataset.test_labels).cuda()\n", 311 | "\n", 312 | "\n", 313 | "output = model(evaluate_x)\n", 314 | "pred = output.data.max(1)[1]\n", 315 | "d = pred.eq(evaluate_y.data).cpu()\n", 316 | "accuracy = d.sum()/d.size()[0]\n", 317 | "\n", 318 | "print('Accuracy:', accuracy)" 319 | ] 320 | } 321 | ], 322 | "metadata": { 323 | "kernelspec": { 324 | "display_name": "Python 3", 325 | "language": "python", 326 | "name": "python3" 327 | }, 328 | "language_info": { 329 | "codemirror_mode": { 330 | "name": "ipython", 331 | "version": 3 332 | }, 333 | "file_extension": ".py", 334 | "mimetype": "text/x-python", 335 | "name": "python", 336 | "nbconvert_exporter": "python", 337 | "pygments_lexer": "ipython3", 338 | "version": "3.6.0" 339 | } 340 | }, 341 | "nbformat": 4, 342 | "nbformat_minor": 2 343 | } 344 | -------------------------------------------------------------------------------- /03 [CNN] CIFAR-10 Classification.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 6, 6 | "metadata": { 7 | "collapsed": false 8 | }, 9 | "outputs": [ 10 | { 11 | "name": "stdout", 12 | "output_type": "stream", 13 | "text": [ 14 | "Populating the interactive namespace from numpy and matplotlib\n" 15 | ] 16 | } 17 | ], 18 | "source": [ 19 | "%pylab inline\n", 20 | "\n", 21 | "import torch\n", 22 | "import torchvision\n", 23 | "\n", 24 | "from torch.utils.data import DataLoader\n", 25 | "from torchvision import transforms\n", 26 | "from torchvision.datasets import CIFAR10\n" 27 | ] 28 | }, 29 | { 30 | "cell_type": "markdown", 31 | "metadata": {}, 32 | "source": [ 33 | "# Configuration" 34 | ] 35 | }, 36 | { 37 | "cell_type": "code", 38 | "execution_count": 53, 39 | "metadata": { 40 | "collapsed": true 41 | }, 42 | "outputs": [], 43 | "source": [ 44 | "BATCH_SIZE = 50" 45 | ] 46 | }, 47 | { 48 | "cell_type": "markdown", 49 | "metadata": {}, 50 | "source": [ 51 | "# CIFAR-10 Data" 52 | ] 53 | }, 54 | { 55 | "cell_type": "code", 56 | "execution_count": 55, 57 | "metadata": { 58 | "collapsed": false 59 | }, 60 | "outputs": [ 61 | { 62 | "name": "stdout", 63 | "output_type": "stream", 64 | "text": [ 65 | "Files already downloaded and verified\n", 66 | "Files already downloaded and verified\n" 67 | ] 68 | } 69 | ], 70 | "source": [ 71 | "cifar_transform = transforms.Compose([\n", 72 | " transforms.ToTensor(),\n", 73 | " transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))\n", 74 | "])\n", 75 | "trainset = CIFAR10('./data', train=True, download=True, transform=cifar_transform)\n", 76 | "testset = CIFAR10('./data', train=False, download=True, transform=cifar_transform)\n", 77 | "\n", 78 | "train_loader = DataLoader(trainset, batch_size=BATCH_SIZE, shuffle=True, num_workers=4)\n", 79 | "test_loader = DataLoader(testset, batch_size=BATCH_SIZE, shuffle=True, num_workers=4)\n", 80 | "\n", 81 | "CLASSES = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')" 82 | ] 83 | }, 84 | { 85 | "cell_type": "code", 86 | "execution_count": 64, 87 | "metadata": { 88 | "collapsed": false 89 | }, 90 | "outputs": [ 91 | { 92 | "name": "stdout", 93 | "output_type": "stream", 94 | "text": [ 95 | "Train Shape: (50000, 32, 32, 3)\n", 96 | "Test Shape: (10000, 32, 32, 3)\n", 97 | "Train Loader Weight Size: torch.Size([50, 3, 32, 32])\n", 98 | "Train Loader Bias Size: torch.Size([50])\n" 99 | ] 100 | } 101 | ], 102 | "source": [ 103 | "print('Train Shape:', trainset.train_data.shape)\n", 104 | "print('Test Shape:', testset.test_data.shape)\n", 105 | "\n", 106 | "print('Train Loader Weight Size:', iter(train_loader).next()[0].size())\n", 107 | "print('Train Loader Bias Size:', iter(train_loader).next()[1].size())\n", 108 | "\n", 109 | "# iter(train_loader).next()[0].numpy().shape()" 110 | ] 111 | }, 112 | { 113 | "cell_type": "code", 114 | "execution_count": 122, 115 | "metadata": { 116 | "collapsed": false 117 | }, 118 | "outputs": [ 119 | { 120 | "data": { 121 | "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAGRCAYAAACNEAwhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMBgMBsOtg90AGQwGg8FguHX4DyR1qMWNiSz6AAAAAElFTkSuQmCC\n", 122 | "text/plain": [ 123 | "" 124 | ] 125 | }, 126 | "metadata": {}, 127 | "output_type": "display_data" 128 | } 129 | ], 130 | "source": [ 131 | "_to_pil = transforms.ToPILImage()\n", 132 | "\n", 133 | "figsize(10, 7)\n", 134 | "fig, subplots = pylab.subplots(5, 7) # subplots(y축, x축 갯수)\n", 135 | "\n", 136 | "idx = 10\n", 137 | "for _subs in subplots:\n", 138 | " for subplot in _subs:\n", 139 | " images, labels = train_iter.next()\n", 140 | " subplot.get_xaxis().set_visible(False)\n", 141 | " subplot.get_yaxis().set_visible(False)\n", 142 | " subplot.imshow(_to_pil(trainset.train_data[idx]))\n", 143 | " idx += 1" 144 | ] 145 | }, 146 | { 147 | "cell_type": "code", 148 | "execution_count": 140, 149 | "metadata": { 150 | "collapsed": false 151 | }, 152 | "outputs": [ 153 | { 154 | "data": { 155 | "text/plain": [ 156 | "array([[[[ 187. , 185. , 181. , ..., 183. ,\n", 157 | " 185. , 185. ],\n", 158 | " [ 223. , 219. , 217. , ..., 227. ,\n", 159 | " 225. , 221. ],\n", 160 | " [ 207. , 209. , 203. , ..., 213. ,\n", 161 | " 211. , 211. ],\n", 162 | " ..., \n", 163 | " [ 131. , 125.00000763, 37.00001144, ..., -34.99999619,\n", 164 | " 5.00001478, 113.00000763],\n", 165 | " [ 117.00000763, 115.00000763, 89.00000763, ..., 3.00001502,\n", 166 | " 5.00001478, 113.00000763],\n", 167 | " [ 135. , 131. , 103.00000763, ..., -78.99999237,\n", 168 | " -28.99999809, 81.00000763]],\n", 169 | "\n", 170 | " [[ 187. , 185. , 181. , ..., 185. ,\n", 171 | " 187. , 187. ],\n", 172 | " [ 229. , 225. , 221. , ..., 233. ,\n", 173 | " 231. , 227. ],\n", 174 | " [ 215. , 215. , 209. , ..., 223. ,\n", 175 | " 219. , 221. ],\n", 176 | " ..., \n", 177 | " [ 167. , 155. , 61.00001144, ..., -12.99999905,\n", 178 | " 27.00001335, 135. ],\n", 179 | " [ 149. , 141. , 109.00000763, ..., 31.00001335,\n", 180 | " 33.00001144, 139. ],\n", 181 | " [ 163. , 155. , 119.00000763, ..., -50.99999619,\n", 182 | " 1.00001514, 107.00000763]],\n", 183 | "\n", 184 | " [[ 207. , 205. , 201. , ..., 189. ,\n", 185 | " 191. , 191. ],\n", 186 | " [ 245. , 243. , 239. , ..., 239. ,\n", 187 | " 237. , 233. ],\n", 188 | " [ 231. , 233. , 227. , ..., 231. ,\n", 189 | " 227. , 229. ],\n", 190 | " ..., \n", 191 | " [ 195. , 177. , 79.00000763, ..., 7.00001478,\n", 192 | " 47.00001144, 155. ],\n", 193 | " [ 179. , 165. , 129. , ..., 51.00001144,\n", 194 | " 55.00001144, 163. ],\n", 195 | " [ 193. , 179. , 143. , ..., -24.99999809,\n", 196 | " 25.00001335, 133. ]]],\n", 197 | "\n", 198 | "\n", 199 | " [[[ 85.00000763, 87.00000763, 63.00001144, ..., 21.00001335,\n", 200 | " -8.99999905, -44.99999619],\n", 201 | " [ 65.00000763, 21.00001335, 15.00001431, ..., 35.00001144,\n", 202 | " -30.99999809, -22.99999809],\n", 203 | " [ 53.00001144, -46.99999619, -4.99999952, ..., 25.00001335,\n", 204 | " 15.00001431, -18.99999809],\n", 205 | " ..., \n", 206 | " [-100.99999237, -90.99999237, -88.99999237, ..., -76.99999237,\n", 207 | " -78.99999237, -84.99999237],\n", 208 | " [ -96.99999237, -90.99999237, -84.99999237, ..., -74.99999237,\n", 209 | " -80.99999237, -76.99999237],\n", 210 | " [ -90.99999237, -82.99999237, -84.99999237, ..., -80.99999237,\n", 211 | " -78.99999237, -66.99999237]],\n", 212 | "\n", 213 | " [[ 87.00000763, 89.00000763, 65.00000763, ..., 29.00001335,\n", 214 | " -0.99999994, -38.99999619],\n", 215 | " [ 67.00000763, 23.00001335, 17.00001335, ..., 43.00001144,\n", 216 | " -22.99999809, -14.99999905],\n", 217 | " [ 55.00001144, -44.99999619, -2.99999976, ..., 33.00001144,\n", 218 | " 23.00001335, -12.99999905],\n", 219 | " ..., \n", 220 | " [ -62.99999619, -62.99999619, -64.99999237, ..., -44.99999619,\n", 221 | " -46.99999619, -52.99999619],\n", 222 | " [ -66.99999237, -60.99999619, -56.99999619, ..., -44.99999619,\n", 223 | " -50.99999619, -46.99999619],\n", 224 | " [ -62.99999619, -54.99999619, -56.99999619, ..., -50.99999619,\n", 225 | " -48.99999619, -36.99999619]],\n", 226 | "\n", 227 | " [[ 75.00000763, 77.00000763, 53.00001144, ..., -22.99999809,\n", 228 | " -48.99999619, -78.99999237],\n", 229 | " [ 55.00001144, 11.00001431, 5.00001478, ..., -8.99999905,\n", 230 | " -70.99999237, -54.99999619],\n", 231 | " [ 43.00001144, -56.99999619, -14.99999905, ..., -18.99999809,\n", 232 | " -24.99999809, -52.99999619],\n", 233 | " ..., \n", 234 | " [-116.99999237, -108.99999237, -102.99999237, ..., -100.99999237,\n", 235 | " -100.99999237, -106.99999237],\n", 236 | " [-114.99999237, -106.99999237, -100.99999237, ..., -100.99999237,\n", 237 | " -106.99999237, -102.99999237],\n", 238 | " [-108.99999237, -100.99999237, -102.99999237, ..., -108.99999237,\n", 239 | " -106.99999237, -94.99999237]]],\n", 240 | "\n", 241 | "\n", 242 | " [[[ -54.99999619, -58.99999619, -54.99999619, ..., -44.99999619,\n", 243 | " -44.99999619, -50.99999619],\n", 244 | " [ -98.99999237, -96.99999237, -96.99999237, ..., -94.99999237,\n", 245 | " -84.99999237, -52.99999619],\n", 246 | " [-139. , -139. , -137. , ..., -112.99999237,\n", 247 | " -108.99999237, -64.99999237],\n", 248 | " ..., \n", 249 | " [ 5.00001478, 1.00001514, -4.99999952, ..., 7.00001478,\n", 250 | " 5.00001478, 3.00001502],\n", 251 | " [ 21.00001335, 3.00001502, -6.99999952, ..., 7.00001478,\n", 252 | " 13.00001431, 23.00001335],\n", 253 | " [ 17.00001335, 1.00001514, -0.99999994, ..., 35.00001144,\n", 254 | " 35.00001144, 47.00001144]],\n", 255 | "\n", 256 | " [[ -50.99999619, -54.99999619, -48.99999619, ..., -30.99999809,\n", 257 | " -34.99999619, -40.99999619],\n", 258 | " [ -88.99999237, -88.99999237, -86.99999237, ..., -66.99999237,\n", 259 | " -62.99999619, -38.99999619],\n", 260 | " [-110.99999237, -112.99999237, -110.99999237, ..., -82.99999237,\n", 261 | " -88.99999237, -52.99999619],\n", 262 | " ..., \n", 263 | " [ -6.99999952, -10.99999905, -16.99999809, ..., -8.99999905,\n", 264 | " -14.99999905, -18.99999809],\n", 265 | " [ 9.00001431, -8.99999905, -18.99999809, ..., -16.99999809,\n", 266 | " -12.99999905, -2.99999976],\n", 267 | " [ 5.00001478, -10.99999905, -12.99999905, ..., 7.00001478,\n", 268 | " 5.00001478, 13.00001431]],\n", 269 | "\n", 270 | " [[ -38.99999619, -42.99999619, -36.99999619, ..., -24.99999809,\n", 271 | " -22.99999809, -24.99999809],\n", 272 | " [ -80.99999237, -80.99999237, -78.99999237, ..., -60.99999619,\n", 273 | " -52.99999619, -24.99999809],\n", 274 | " [-108.99999237, -110.99999237, -108.99999237, ..., -96.99999237,\n", 275 | " -94.99999237, -52.99999619],\n", 276 | " ..., \n", 277 | " [ 1.00001514, -2.99999976, -8.99999905, ..., -18.99999809,\n", 278 | " -20.99999809, -24.99999809],\n", 279 | " [ 17.00001335, -0.99999994, -10.99999905, ..., -18.99999809,\n", 280 | " -12.99999905, -2.99999976],\n", 281 | " [ 13.00001431, -2.99999976, -4.99999952, ..., 9.00001431,\n", 282 | " 9.00001431, 17.00001335]]],\n", 283 | "\n", 284 | "\n", 285 | " ..., \n", 286 | " [[[ -90.99999237, -96.99999237, -96.99999237, ..., -44.99999619,\n", 287 | " -50.99999619, -16.99999809],\n", 288 | " [ -94.99999237, -88.99999237, -86.99999237, ..., -64.99999237,\n", 289 | " -26.99999809, 13.00001431],\n", 290 | " [ -82.99999237, -78.99999237, -100.99999237, ..., -36.99999619,\n", 291 | " -30.99999809, 1.00001514],\n", 292 | " ..., \n", 293 | " [ 1.00001514, -12.99999905, -54.99999619, ..., -34.99999619,\n", 294 | " -78.99999237, -98.99999237],\n", 295 | " [ 49.00001144, 23.00001335, 1.00001514, ..., -100.99999237,\n", 296 | " -112.99999237, -94.99999237],\n", 297 | " [ 23.00001335, 7.00001478, 41.00001144, ..., -98.99999237,\n", 298 | " -124.99999237, -88.99999237]],\n", 299 | "\n", 300 | " [[-126.99999237, -126.99999237, -122.99999237, ..., -70.99999237,\n", 301 | " -70.99999237, -24.99999809],\n", 302 | " [-131. , -110.99999237, -96.99999237, ..., -72.99999237,\n", 303 | " -28.99999809, 19.00001335],\n", 304 | " [-122.99999237, -106.99999237, -116.99999237, ..., -54.99999619,\n", 305 | " -46.99999619, -12.99999905],\n", 306 | " ..., \n", 307 | " [ 13.00001431, 3.00001502, -32.99999619, ..., -18.99999809,\n", 308 | " -66.99999237, -90.99999237],\n", 309 | " [ 49.00001144, 29.00001335, 15.00001431, ..., -86.99999237,\n", 310 | " -98.99999237, -80.99999237],\n", 311 | " [ 27.00001335, 19.00001335, 59.00001144, ..., -96.99999237,\n", 312 | " -124.99999237, -80.99999237]],\n", 313 | "\n", 314 | " [[-181. , -191. , -185. , ..., -110.99999237,\n", 315 | " -112.99999237, -72.99999237],\n", 316 | " [-189. , -183. , -169. , ..., -122.99999237,\n", 317 | " -80.99999237, -36.99999619],\n", 318 | " [-177. , -171. , -183. , ..., -104.99999237,\n", 319 | " -96.99999237, -64.99999237],\n", 320 | " ..., \n", 321 | " [ -86.99999237, -94.99999237, -153. , ..., -124.99999237,\n", 322 | " -161. , -177. ],\n", 323 | " [ -26.99999809, -52.99999619, -86.99999237, ..., -185. ,\n", 324 | " -183. , -165. ],\n", 325 | " [ -56.99999619, -78.99999237, -52.99999619, ..., -175. ,\n", 326 | " -191. , -163. ]]],\n", 327 | "\n", 328 | "\n", 329 | " [[[-112.99999237, -78.99999237, -90.99999237, ..., -197. ,\n", 330 | " -199. , -201. ],\n", 331 | " [-187. , -179. , -179. , ..., -199. ,\n", 332 | " -199. , -199. ],\n", 333 | " [-199. , -199. , -195. , ..., -197. ,\n", 334 | " -199. , -201. ],\n", 335 | " ..., \n", 336 | " [ -92.99999237, -106.99999237, -118.99999237, ..., -78.99999237,\n", 337 | " -98.99999237, -98.99999237],\n", 338 | " [ -86.99999237, -100.99999237, -114.99999237, ..., -58.99999619,\n", 339 | " -74.99999237, -86.99999237],\n", 340 | " [ -82.99999237, -86.99999237, -92.99999237, ..., -58.99999619,\n", 341 | " -68.99999237, -84.99999237]],\n", 342 | "\n", 343 | " [[-102.99999237, -68.99999237, -80.99999237, ..., -185. ,\n", 344 | " -189. , -191. ],\n", 345 | " [-175. , -169. , -169. , ..., -187. ,\n", 346 | " -189. , -189. ],\n", 347 | " [-189. , -189. , -185. , ..., -185. ,\n", 348 | " -189. , -191. ],\n", 349 | " ..., \n", 350 | " [ -80.99999237, -88.99999237, -94.99999237, ..., -64.99999237,\n", 351 | " -76.99999237, -74.99999237],\n", 352 | " [ -84.99999237, -90.99999237, -98.99999237, ..., -48.99999619,\n", 353 | " -58.99999619, -70.99999237],\n", 354 | " [ -86.99999237, -86.99999237, -84.99999237, ..., -52.99999619,\n", 355 | " -60.99999619, -76.99999237]],\n", 356 | "\n", 357 | " [[ -90.99999237, -56.99999619, -68.99999237, ..., -173. ,\n", 358 | " -179. , -179. ],\n", 359 | " [-163. , -157. , -157. , ..., -173. ,\n", 360 | " -177. , -177. ],\n", 361 | " [-177. , -177. , -173. , ..., -171. ,\n", 362 | " -177. , -179. ],\n", 363 | " ..., \n", 364 | " [-114.99999237, -122.99999237, -131. , ..., -110.99999237,\n", 365 | " -122.99999237, -124.99999237],\n", 366 | " [-120.99999237, -129. , -137. , ..., -92.99999237,\n", 367 | " -104.99999237, -116.99999237],\n", 368 | " [-126.99999237, -126.99999237, -129. , ..., -94.99999237,\n", 369 | " -102.99999237, -120.99999237]]],\n", 370 | "\n", 371 | "\n", 372 | " [[[ 237. , 247. , 249. , ..., 55.00001144,\n", 373 | " 63.00001144, 65.00000763],\n", 374 | " [ 245. , 255. , 255. , ..., 31.00001335,\n", 375 | " 33.00001144, 41.00001144],\n", 376 | " [ 245. , 253. , 253. , ..., 25.00001335,\n", 377 | " 27.00001335, 41.00001144],\n", 378 | " ..., \n", 379 | " [ -2.99999976, -8.99999905, -4.99999952, ..., -106.99999237,\n", 380 | " -102.99999237, -114.99999237],\n", 381 | " [ -16.99999809, -14.99999905, -10.99999905, ..., -110.99999237,\n", 382 | " -112.99999237, -126.99999237],\n", 383 | " [ 9.00001431, 17.00001335, 27.00001335, ..., -62.99999619,\n", 384 | " -64.99999237, -76.99999237]],\n", 385 | "\n", 386 | " [[ 249. , 247. , 247. , ..., 133. ,\n", 387 | " 131. , 133. ],\n", 388 | " [ 251. , 255. , 255. , ..., 121.00000763,\n", 389 | " 123.00000763, 125.00000763],\n", 390 | " [ 247. , 251. , 251. , ..., 125.00000763,\n", 391 | " 131. , 137. ],\n", 392 | " ..., \n", 393 | " [ -10.99999905, -14.99999905, -14.99999905, ..., -143. ,\n", 394 | " -145. , -151. ],\n", 395 | " [ -24.99999809, -20.99999809, -16.99999809, ..., -145. ,\n", 396 | " -151. , -161. ],\n", 397 | " [ 3.00001502, 13.00001431, 21.00001335, ..., -90.99999237,\n", 398 | " -92.99999237, -100.99999237]],\n", 399 | "\n", 400 | " [[ 251. , 247. , 245. , ..., 235. ,\n", 401 | " 235. , 225. ],\n", 402 | " [ 255. , 251. , 251. , ..., 249. ,\n", 403 | " 251. , 233. ],\n", 404 | " [ 249. , 245. , 247. , ..., 247. ,\n", 405 | " 251. , 239. ],\n", 406 | " ..., \n", 407 | " [ -54.99999619, -72.99999237, -62.99999619, ..., -175. ,\n", 408 | " -175. , -179. ],\n", 409 | " [ -60.99999619, -66.99999237, -56.99999619, ..., -179. ,\n", 410 | " -183. , -189. ],\n", 411 | " [ -24.99999809, -18.99999809, -4.99999952, ..., -118.99999237,\n", 412 | " -120.99999237, -126.99999237]]]], dtype=float32)" 413 | ] 414 | }, 415 | "execution_count": 140, 416 | "metadata": {}, 417 | "output_type": "execute_result" 418 | } 419 | ], 420 | "source": [ 421 | "transforms.Scale()\n", 422 | "iter(train_loader).next()[0][0].numpy()" 423 | ] 424 | }, 425 | { 426 | "cell_type": "code", 427 | "execution_count": null, 428 | "metadata": { 429 | "collapsed": true 430 | }, 431 | "outputs": [], 432 | "source": [] 433 | } 434 | ], 435 | "metadata": { 436 | "kernelspec": { 437 | "display_name": "Python 3", 438 | "language": "python", 439 | "name": "python3" 440 | }, 441 | "language_info": { 442 | "codemirror_mode": { 443 | "name": "ipython", 444 | "version": 3 445 | }, 446 | "file_extension": ".py", 447 | "mimetype": "text/x-python", 448 | "name": "python", 449 | "nbconvert_exporter": "python", 450 | "pygments_lexer": "ipython3", 451 | "version": "3.6.0" 452 | } 453 | }, 454 | "nbformat": 4, 455 | "nbformat_minor": 2 456 | } 457 | --------------------------------------------------------------------------------