training_a_classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![alt text](pictures/cifar10.png \"CIFAR10\")\n",
    "# Training a Image Classifier\n",
    "### New Package in this notes: <span style=\"color:red\">torchvision</span>\n",
    "http://pytorch.org/docs/torchvision/torchvision.html\n",
    "#### Traing a image classifier based on CIFAR10 as the following steps:\n",
    "- Load and normalizing the CIFAR10 training and test datasets using torchvision\n",
    "- Define a Convolution Neural Network\n",
    "- Define a loss function\n",
    "- Train the network on the training data\n",
    "- Test the network on the test data\n",
    "---\n",
    "## 1. Loading and normalizing CIFAR10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "\n",
    "transform = transforms.Compose(\n",
    "    [transforms.ToTensor(),\n",
    "     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])\n",
    "\n",
    "trainset = torchvision.datasets.CIFAR10(root='./data', train=True,\n",
    "                                        download=True, transform=transform)\n",
    "trainloader = torch.utils.data.DataLoader(trainset, batch_size=4,\n",
    "                                          shuffle=True, num_workers=2)\n",
    "\n",
    "testset = torchvision.datasets.CIFAR10(root='./data', train=False,\n",
    "                                       download=True, transform=transform)\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=4,\n",
    "                                         shuffle=False, num_workers=2)\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat',\n",
    "           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "上面代码第一次运行时显示：\n",
    "```\n",
    "Downloading http://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to ./data/cifar-10-python.tar.gz\n",
    "Files already downloaded and verified\n",
    "```\n",
    "---\n",
    "### 使用torchvision package下载并处理数据集。\n",
    "- **torchvision.transforms.Compose()**可以把多个transforms.\\*()函数集合在一起。这些transforms.\\*()把输入图像转换成某种格式。例如上面的代码中把transform.ToTensor()和transform.Normalize()这两个函数组合了在一起。\n",
    " - **transforms.ToTensor()**: Converts a PIL.Image or numpy.ndarray (H x W x C) in the range [0, 255] to a torch.FloatTensor of shape (C x H x W) in the range [0.0, 1.0].\n",
    " - **transforms.Normalize(mean, std)**: Given mean: (R, G, B) and std: (R, G, B), will normalize each channel of the torch.\\*Tensor, i.e. channel = (channel - mean) / std. 也就是标准化图像数据。\n",
    "- **torchvision.datasets.CIFAR10()**：\n",
    " - root：数据集的根目录\n",
    " - train：=True时为训练集，=False时为测试集\n",
    " - download：=True时表示从网上下载数据集到根目录下，如果已经下载过，则什么都不做。\n",
    " - transform: 指明需要的转换步骤。\n",
    "- **torch.utils.data.DataLoader(dataset, batch_size=1, shuffle=False, sampler=None, num_workers=0, collate_fn=function default_collate, pin_memory=False, drop_last=False)**: 数据加载程序\n",
    " - dataset(trainset/testset)：从哪里加载数据\n",
    " - batch_size: 每一个batch中有几个样本\n",
    " - shuffle：如果时True，则在每一个epoch后打乱数据的顺序\n",
    " - num_workers: 使用几个线程载入数据。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<torchvision.datasets.cifar.CIFAR10 object at 0x7f8710325310>\n",
      "<torch.utils.data.dataloader.DataLoader object at 0x7f87103ea3d0>\n"
     ]
    }
   ],
   "source": [
    "print(trainset)\n",
    "print(trainloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  car  deer plane  bird\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAB6CAYAAACvHqiXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd110bdda90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "# functions to show an image\n",
    "\n",
    "def imshow(img):\n",
    "    img = img / 2 + 0.5     # unnormalize\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "\n",
    "# get some random training images\n",
    "dataiter = iter(trainloader)     # iter() is to read lines of a file until a certain line is reached\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "# show images\n",
    "imshow(torchvision.utils.make_grid(images))\n",
    "# print labels\n",
    "print(' '.join('%5s' % classes[labels[j]] for j in range(4)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. 定义卷积神经网络"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from torch.autograd import Variable\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(3, 6, 5)\n",
    "        self.pool = nn.MaxPool2d(2, 2)\n",
    "        self.conv2 = nn.Conv2d(6, 16, 5)\n",
    "        self.fc1 = nn.Linear(16 * 5 * 5, 120)\n",
    "        self.fc2 = nn.Linear(120, 84)\n",
    "        self.fc3 = nn.Linear(84, 10)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        x = x.view(-1, 16 * 5 * 5)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = self.fc3(x)\n",
    "        return x\n",
    "\n",
    "net = Net()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "注：\n",
    "```python\n",
    ".view(): Returns a new tensor with the same data but different size.\n",
    "E.g.\n",
    ">>> x = torch.randn(4, 4)\n",
    ">>> x.size()\n",
    "torch.Size([4, 4])\n",
    ">>> y = x.view(16)\n",
    ">>> y.size()\n",
    "torch.Size([16])\n",
    ">>> z = x.view(-1, 8)  # the size -1 is inferred from other dimensions\n",
    ">>> z.size()\n",
    "torch.Size([2, 8])\n",
    "```\n",
    "\n",
    "## 3. 定义损失函数和优化器"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(net.parameters(), lr = 0.001, momentum=0.9)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. 训练网络"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1,  2000] loss: 2.253\n",
      "[1,  4000] loss: 1.857\n",
      "[1,  6000] loss: 1.665\n",
      "[1,  8000] loss: 1.578\n",
      "[1, 10000] loss: 1.514\n",
      "[1, 12000] loss: 1.456\n",
      "[2,  2000] loss: 1.397\n",
      "[2,  4000] loss: 1.375\n",
      "[2,  6000] loss: 1.353\n",
      "[2,  8000] loss: 1.330\n",
      "[2, 10000] loss: 1.304\n",
      "[2, 12000] loss: 1.313\n",
      "Finished Training\n"
     ]
    }
   ],
   "source": [
    "for epoch in range(2):\n",
    "    running_loss = 0\n",
    "    for i, data in enumerate(trainloader, 0):   # enumerate(thing) return (0, thing[0]), (1, thing[1]),...\n",
    "        # get the inputs\n",
    "        inputs, labels = data\n",
    "        \n",
    "        # wrap them in Variable\n",
    "        inputs, labels = Variable(inputs), Variable(labels)\n",
    "        \n",
    "        # zero the parameter gradients\n",
    "        optimizer.zero_grad()\n",
    "        \n",
    "        # forward + backward + optimize\n",
    "        outputs = net(inputs)\n",
    "        loss = criterion(outputs, labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        # print\n",
    "        running_loss += loss.data[0]\n",
    "        if i % 2000 == 1999:                  # print every 2000 mini-batch\n",
    "            print('[%d, %5d] loss: %.3f' %\n",
    "                 (epoch + 1, i + 1, running_loss / 2000))\n",
    "            running_loss = 0\n",
    "            \n",
    "print('Finished Training')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. 在测试数据集上测试神经网络\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " 3\n",
      " 8\n",
      " 8\n",
      " 0\n",
      "[torch.LongTensor of size 4]\n",
      "\n",
      "('Ground Truth: ', '  cat  ship  ship plane')\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAB6CAYAAACvHqiXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+P/jyitvLuJHGwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd0da217990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataiter = iter(testloader)\n",
    "images, labels = dataiter.next()\n",
    "print(labels)\n",
    "\n",
    "imshow(torchvision.utils.make_grid(images))\n",
    "print('Ground Truth: ', ' '.join('%5s' % classes[labels[j]] for j in range(4)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Variable containing:\n",
      "\n",
      "Columns 0 to 7 \n",
      " -2.3651  -1.0204  -0.2224   3.3335  -0.0557   2.2254   0.8923  -1.1199\n",
      "  7.1263   6.5507  -2.7360  -3.8514  -4.0859  -6.2206  -6.7404  -5.4120\n",
      "  2.2768   2.2120  -0.5086  -1.3173  -1.2032  -2.6589  -2.9600  -1.7045\n",
      "  4.9461   2.4030  -0.7406  -2.3374  -1.7571  -4.3366  -4.0236  -3.4177\n",
      "\n",
      "Columns 8 to 9 \n",
      " -0.4827  -1.7411\n",
      " 10.3903   3.7646\n",
      "  3.8142   1.4575\n",
      "  6.9897   1.8331\n",
      "[torch.FloatTensor of size 4x10]\n",
      "\n",
      "\n",
      "\n",
      "Columns 0 to 7 \n",
      " -2.3651  -1.0204  -0.2224   3.3335  -0.0557   2.2254   0.8923  -1.1199\n",
      "  7.1263   6.5507  -2.7360  -3.8514  -4.0859  -6.2206  -6.7404  -5.4120\n",
      "  2.2768   2.2120  -0.5086  -1.3173  -1.2032  -2.6589  -2.9600  -1.7045\n",
      "  4.9461   2.4030  -0.7406  -2.3374  -1.7571  -4.3366  -4.0236  -3.4177\n",
      "\n",
      "Columns 8 to 9 \n",
      " -0.4827  -1.7411\n",
      " 10.3903   3.7646\n",
      "  3.8142   1.4575\n",
      "  6.9897   1.8331\n",
      "[torch.FloatTensor of size 4x10]\n",
      "\n",
      "\n",
      "  3.3335\n",
      " 10.3903\n",
      "  3.8142\n",
      "  6.9897\n",
      "[torch.FloatTensor of size 4x1]\n",
      "\n",
      "\n",
      " 3\n",
      " 8\n",
      " 8\n",
      " 8\n",
      "[torch.LongTensor of size 4x1]\n",
      "\n",
      "('Predicted: ', '  cat  ship  ship  ship')\n"
     ]
    }
   ],
   "source": [
    "outputs = net(Variable(images))\n",
    "print(outputs)\n",
    "print(outputs.data)\n",
    "_, predicted = torch.max(outputs.data, 1)\n",
    "print(_)\n",
    "print(predicted)\n",
    "print('Predicted: ', \" \".join('%5s' % classes[predicted[j][0]] for j in range(4)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- torch.max()函数返回数据数据的最大值，参数1表示在哪个维度检测最大值。上面的代码中，变量\\_是最大值，变量predicted返回最大值的索引值，也就是最大值的位置。\n",
    "- classes是在cell 1 中定义的类别标签。\n",
    "- 上面代码最后一行predicted[j][0]就是访问predicted的地j行第0列的元素，因为predicted是4x1的，所以写成predicted[j]也是一样可以的。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the 10000 test images: 54 %\n"
     ]
    }
   ],
   "source": [
    "correct = 0\n",
    "total = 0\n",
    "for data in testloader:\n",
    "    images, labels = data\n",
    "    outputs = net(Variable(images))\n",
    "    _, predicted = torch.max(outputs.data, 1)\n",
    "    total += labels.size(0)\n",
    "    correct += (predicted == labels).sum()\n",
    "print('Accuracy of the network on the 10000 test images: %d %%' % (100 * correct / total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of plane : 58 %\n",
      "Accuracy of   car : 65 %\n",
      "Accuracy of  bird : 37 %\n",
      "Accuracy of   cat : 51 %\n",
      "Accuracy of  deer : 28 %\n",
      "Accuracy of   dog : 48 %\n",
      "Accuracy of  frog : 62 %\n",
      "Accuracy of horse : 62 %\n",
      "Accuracy of  ship : 80 %\n",
      "Accuracy of truck : 42 %\n"
     ]
    }
   ],
   "source": [
    "class_correct = list(0. for i in range(10))\n",
    "class_total = list(0. for i in range(10))\n",
    "for data in testloader:\n",
    "    iamges, labels = data\n",
    "    outputs = net(Variable(iamges))\n",
    "    _, predicted = torch.max(outputs.data, 1)\n",
    "    c = (predicted == labels).squeeze()\n",
    "    for i in range(4):\n",
    "        label = labels[i]\n",
    "        class_correct[label] += c[i]\n",
    "        class_total[label] += 1\n",
    "\n",
    "\n",
    "for i in range(10):\n",
    "    print('Accuracy of %5s : %2d %%' % (\n",
    "        classes[i], 100 * class_correct[i] / class_total[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "AssertionError",
     "evalue": "\nFound no NVIDIA driver on your system. Please check that you\nhave an NVIDIA GPU and installed a driver from\nhttp://www.nvidia.com/Download/index.aspx",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-19-7b748db2b046>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mnet\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/nn/modules/module.pyc\u001b[0m in \u001b[0;36mcuda\u001b[0;34m(self, device_id)\u001b[0m\n\u001b[1;32m    145\u001b[0m                 \u001b[0mcopied\u001b[0m \u001b[0mto\u001b[0m \u001b[0mthat\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    146\u001b[0m         \"\"\"\n\u001b[0;32m--> 147\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice_id\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    148\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    149\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mcpu\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice_id\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/nn/modules/module.pyc\u001b[0m in \u001b[0;36m_apply\u001b[0;34m(self, fn)\u001b[0m\n\u001b[1;32m    116\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    117\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mmodule\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mchildren\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 118\u001b[0;31m             \u001b[0mmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mparam\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parameters\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/nn/modules/module.pyc\u001b[0m in \u001b[0;36m_apply\u001b[0;34m(self, fn)\u001b[0m\n\u001b[1;32m    122\u001b[0m                 \u001b[0;31m# Variables stored in modules are graph leaves, and we don't\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    123\u001b[0m                 \u001b[0;31m# want to create copy nodes, so we have to unpack the data.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 124\u001b[0;31m                 \u001b[0mparam\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparam\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    125\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mparam\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_grad\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    126\u001b[0m                     \u001b[0mparam\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_grad\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparam\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_grad\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/nn/modules/module.pyc\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(t)\u001b[0m\n\u001b[1;32m    145\u001b[0m                 \u001b[0mcopied\u001b[0m \u001b[0mto\u001b[0m \u001b[0mthat\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    146\u001b[0m         \"\"\"\n\u001b[0;32m--> 147\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice_id\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    148\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    149\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mcpu\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice_id\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/_utils.pyc\u001b[0m in \u001b[0;36m_cuda\u001b[0;34m(self, device, async)\u001b[0m\n\u001b[1;32m     63\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     64\u001b[0m             \u001b[0mnew_type\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mgetattr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__class__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 65\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mnew_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msize\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0masync\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     66\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     67\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/cuda/__init__.pyc\u001b[0m in \u001b[0;36m__new__\u001b[0;34m(cls, *args, **kwargs)\u001b[0m\n\u001b[1;32m    270\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    271\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m__new__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcls\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 272\u001b[0;31m         \u001b[0m_lazy_init\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    273\u001b[0m         \u001b[0;31m# We need this method only for lazy init, so we can remove it\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    274\u001b[0m         \u001b[0;32mdel\u001b[0m \u001b[0m_CudaBase\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__new__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/cuda/__init__.pyc\u001b[0m in \u001b[0;36m_lazy_init\u001b[0;34m()\u001b[0m\n\u001b[1;32m     82\u001b[0m         raise RuntimeError(\n\u001b[1;32m     83\u001b[0m             \"Cannot re-initialize CUDA in forked subprocess. \" + msg)\n\u001b[0;32m---> 84\u001b[0;31m     \u001b[0m_check_driver\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     85\u001b[0m     \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_C\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_cuda_init\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     86\u001b[0m     \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_C\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_cuda_sparse_init\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/zxw/anaconda2/lib/python2.7/site-packages/torch/cuda/__init__.pyc\u001b[0m in \u001b[0;36m_check_driver\u001b[0;34m()\u001b[0m\n\u001b[1;32m     56\u001b[0m \u001b[0mFound\u001b[0m \u001b[0mno\u001b[0m \u001b[0mNVIDIA\u001b[0m \u001b[0mdriver\u001b[0m \u001b[0mon\u001b[0m \u001b[0myour\u001b[0m \u001b[0msystem\u001b[0m\u001b[0;34m.\u001b[0m \u001b[0mPlease\u001b[0m \u001b[0mcheck\u001b[0m \u001b[0mthat\u001b[0m \u001b[0myou\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     57\u001b[0m \u001b[0mhave\u001b[0m \u001b[0man\u001b[0m \u001b[0mNVIDIA\u001b[0m \u001b[0mGPU\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0minstalled\u001b[0m \u001b[0ma\u001b[0m \u001b[0mdriver\u001b[0m \u001b[0;32mfrom\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 58\u001b[0;31m http://www.nvidia.com/Download/index.aspx\"\"\")\n\u001b[0m\u001b[1;32m     59\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     60\u001b[0m             \u001b[0;31m# TODO: directly link to the alternative bin that needs install\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAssertionError\u001b[0m: \nFound no NVIDIA driver on your system. Please check that you\nhave an NVIDIA GPU and installed a driver from\nhttp://www.nvidia.com/Download/index.aspx"
     ]
    }
   ],
   "source": [
    "net.cuda()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}