DL-with-Python-and-PyTorch2/pytorch-12/pytorch-12-02.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 12.4 微调实例\n",
    "\t微调允许修改预先训练好的网络参数来学习目标任务，所以，微调的训练时间要比特征抽取的训练时间长，但精度更高。微调的大致过程是在预先训练过的网络上添加新的随机初始化层，也会更新预先训练的网络参数，但会使用较小的学习率以防止预先训练好的参数发生较大改变。\n",
    "\t常用的方法是固定底层的参数，调整一些顶层或具体层的参数。这样做的好处是可以减少训练参数的数量，也可以克服过拟合现象的发生。尤其是在目标任务的数据量不足够大时，该方法的实践效果更好。实际上，微调要优于特征提取，因为它能够对迁移过来的预训练网络参数进行优化，使其更加适合新的任务。\n",
    "### 12.4.1 数据预处理\n",
    "\t这里对训练数据添加了几种数据增强方法，如图像裁剪、旋转、颜色改变等方法。测试数据与特征提取的测试数据一样，没有变化。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch import nn\n",
    "import torch.nn.functional as F\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "from torchvision import models\n",
    "from torchvision.datasets import ImageFolder\n",
    "from datetime import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "trans_train = transforms.Compose(\n",
    "    [transforms.RandomResizedCrop(size=256, scale=(0.8, 1.0)),\n",
    "     transforms.RandomRotation(degrees=15),\n",
    "     transforms.ColorJitter(),\n",
    "     transforms.RandomResizedCrop(224),\n",
    "     transforms.RandomHorizontalFlip(),     \n",
    "     transforms.ToTensor(),\n",
    "     transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
    "                                 std=[0.229, 0.224, 0.225])])\n",
    "\n",
    "trans_valid = transforms.Compose(\n",
    "    [transforms.Resize(256),\n",
    "    transforms.CenterCrop(224),\n",
    "    transforms.ToTensor(),\n",
    "     transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
    "                                 std=[0.229, 0.224, 0.225])])\n",
    "\n",
    "trainset = torchvision.datasets.CIFAR10(root='../data', train=True,\n",
    "                                        download=False, transform=trans_train)\n",
    "trainloader = torch.utils.data.DataLoader(trainset, batch_size=64,\n",
    "                                          shuffle=True, num_workers=2)\n",
    "\n",
    "testset = torchvision.datasets.CIFAR10(root='../data', train=False,\n",
    "                                       download=False, transform=trans_valid)\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=64,\n",
    "                                         shuffle=False, num_workers=2)\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat',\n",
    "           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  car  deer  ship plane\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "%matplotlib inline\n",
    "\n",
    "# 显示图像\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",
    "    plt.show()\n",
    "\n",
    "\n",
    "# 随机获取部分训练数据\n",
    "dataiter = iter(trainloader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "\n",
    "# 显示图像\n",
    "imshow(torchvision.utils.make_grid(images))\n",
    "# 打印标签\n",
    "print(' '.join('%5s' % classes[labels[j]] for j in range(4)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### 12.4.2 加载预训练模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 使用预训练的模型\n",
    "net = models.resnet18(pretrained=True)\n",
    "#print(net)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ResNet(\n",
      "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
      "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (relu): ReLU(inplace=True)\n",
      "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
      "  (layer1): Sequential(\n",
      "    (0): BasicBlock(\n",
      "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    )\n",
      "    (1): BasicBlock(\n",
      "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    )\n",
      "  )\n",
      "  (layer2): Sequential(\n",
      "    (0): BasicBlock(\n",
      "      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): BasicBlock(\n",
      "      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    )\n",
      "  )\n",
      "  (layer3): Sequential(\n",
      "    (0): BasicBlock(\n",
      "      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): BasicBlock(\n",
      "      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    )\n",
      "  )\n",
      "  (layer4): Sequential(\n",
      "    (0): BasicBlock(\n",
      "      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): BasicBlock(\n",
      "      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    )\n",
      "  )\n",
      "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
      "  (fc): Linear(in_features=512, out_features=1000, bias=True)\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "print(net)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 12.4.3 修改分类器\n",
    "\t修改最后全连接层，把类别数由原来的1000改为10。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ResNet(\n",
       "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (relu): ReLU(inplace=True)\n",
       "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (layer1): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer2): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer3): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer4): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  (fc): Linear(in_features=512, out_features=10, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 将最后的全连接层改成十分类\n",
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
    "net.fc = nn.Linear(512, 10)\n",
    "#net = torch.nn.DataParallel(net)\n",
    "net.to(device)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.cuda.FloatTensor\n"
     ]
    }
   ],
   "source": [
    "# 打出第一层的权重\n",
    "print(net.conv1.weight.type())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### 12.4.4 选择损失函数及优化器\n",
    "\t这里学习率为le-3，使用微调方法训练模型时，一般选择一个稍大一点的学习率，如果选择的学习率太小，效果要差一些。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_acc(output, label):\n",
    "    total = output.shape[0]\n",
    "    _, pred_label = output.max(1)\n",
    "    num_correct = (pred_label == label).sum().item()\n",
    "    return num_correct / total"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.SGD(net.parameters(), lr=1e-3, weight_decay=1e-3,momentum=0.9)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### 12.4.5 训练及验证模型\n",
    "\t训练及验证模型的代码如下："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(net, train_data, valid_data, num_epochs, optimizer, criterion):\n",
    "   \n",
    "    prev_time = datetime.now()\n",
    "    for epoch in range(num_epochs):\n",
    "        train_loss = 0\n",
    "        train_acc = 0\n",
    "        net = net.train()\n",
    "        for im, label in train_data:\n",
    "            im = im.to(device)  # (bs, 3, h, w)\n",
    "            label = label.to(device) # (bs, h, w)\n",
    "            # forward\n",
    "            output = net(im)\n",
    "            loss = criterion(output, label)\n",
    "            # backward\n",
    "            optimizer.zero_grad()\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "\n",
    "            train_loss += loss.item()\n",
    "            train_acc += get_acc(output, label)\n",
    "\n",
    "        cur_time = datetime.now()\n",
    "        h, remainder = divmod((cur_time - prev_time).seconds, 3600)\n",
    "        m, s = divmod(remainder, 60)\n",
    "        time_str = \"Time %02d:%02d:%02d\" % (h, m, s)\n",
    "        if valid_data is not None:\n",
    "            valid_loss = 0\n",
    "            valid_acc = 0\n",
    "            net = net.eval()\n",
    "            for im, label in valid_data:\n",
    "                im = im.to(device)  # (bs, 3, h, w)\n",
    "                label = label.to(device) # (bs, h, w)\n",
    "                output = net(im)\n",
    "                loss = criterion(output, label)\n",
    "                valid_loss += loss.item()\n",
    "                valid_acc += get_acc(output, label)\n",
    "            epoch_str = (\n",
    "                \"Epoch %d. Train Loss: %f, Train Acc: %f, Valid Loss: %f, Valid Acc: %f, \"\n",
    "                % (epoch, train_loss / len(train_data),\n",
    "                   train_acc / len(train_data), valid_loss / len(valid_data),\n",
    "                   valid_acc / len(valid_data)))\n",
    "        else:\n",
    "            epoch_str = (\"Epoch %d. Train Loss: %f, Train Acc: %f, \" %\n",
    "                         (epoch, train_loss / len(train_data),\n",
    "                          train_acc / len(train_data)))\n",
    "        prev_time = cur_time\n",
    "        print(epoch_str + time_str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0. Train Loss: 1.046640, Train Acc: 0.634691, Valid Loss: 0.355521, Valid Acc: 0.880275, Time 00:02:12\n",
      "Epoch 1. Train Loss: 0.689994, Train Acc: 0.758492, Valid Loss: 0.280970, Valid Acc: 0.904757, Time 00:02:25\n",
      "Epoch 2. Train Loss: 0.621406, Train Acc: 0.781630, Valid Loss: 0.248424, Valid Acc: 0.913018, Time 00:02:27\n",
      "Epoch 3. Train Loss: 0.571020, Train Acc: 0.801570, Valid Loss: 0.220384, Valid Acc: 0.925458, Time 00:02:29\n",
      "Epoch 4. Train Loss: 0.540538, Train Acc: 0.812580, Valid Loss: 0.192789, Valid Acc: 0.936007, Time 00:02:29\n",
      "Epoch 5. Train Loss: 0.511390, Train Acc: 0.822530, Valid Loss: 0.182821, Valid Acc: 0.936206, Time 00:02:27\n",
      "Epoch 6. Train Loss: 0.490604, Train Acc: 0.829204, Valid Loss: 0.182214, Valid Acc: 0.936505, Time 00:02:30\n",
      "Epoch 7. Train Loss: 0.480625, Train Acc: 0.831142, Valid Loss: 0.173486, Valid Acc: 0.937799, Time 00:02:30\n",
      "Epoch 8. Train Loss: 0.468967, Train Acc: 0.838175, Valid Loss: 0.167395, Valid Acc: 0.941481, Time 00:02:28\n",
      "Epoch 9. Train Loss: 0.454750, Train Acc: 0.842311, Valid Loss: 0.167120, Valid Acc: 0.943073, Time 00:02:26\n",
      "Epoch 10. Train Loss: 0.444690, Train Acc: 0.845368, Valid Loss: 0.168414, Valid Acc: 0.942576, Time 00:02:26\n",
      "Epoch 11. Train Loss: 0.428961, Train Acc: 0.850384, Valid Loss: 0.159228, Valid Acc: 0.942377, Time 00:02:27\n",
      "Epoch 12. Train Loss: 0.425348, Train Acc: 0.851982, Valid Loss: 0.152707, Valid Acc: 0.946258, Time 00:02:26\n",
      "Epoch 13. Train Loss: 0.414132, Train Acc: 0.855459, Valid Loss: 0.151967, Valid Acc: 0.947253, Time 00:02:26\n",
      "Epoch 14. Train Loss: 0.410402, Train Acc: 0.856298, Valid Loss: 0.152984, Valid Acc: 0.948646, Time 00:02:26\n",
      "Epoch 15. Train Loss: 0.402335, Train Acc: 0.861153, Valid Loss: 0.154637, Valid Acc: 0.947353, Time 00:02:27\n",
      "Epoch 16. Train Loss: 0.391543, Train Acc: 0.863811, Valid Loss: 0.151976, Valid Acc: 0.949045, Time 00:02:26\n",
      "Epoch 17. Train Loss: 0.390811, Train Acc: 0.864390, Valid Loss: 0.157535, Valid Acc: 0.945661, Time 00:02:27\n",
      "Epoch 18. Train Loss: 0.378429, Train Acc: 0.867907, Valid Loss: 0.149089, Valid Acc: 0.951632, Time 00:02:26\n",
      "Epoch 19. Train Loss: 0.374837, Train Acc: 0.869206, Valid Loss: 0.154984, Valid Acc: 0.947253, Time 00:02:26\n"
     ]
    }
   ],
   "source": [
    "train(net, trainloader, testloader, 20, optimizer, criterion)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "由结果可知，微调的训练时间明显大于特征提取的训练时间，其一个循环需要9分钟左右，但验证准确率高达95%。主要这里只循环了20次，如果增加循环次数，准确率应该还可再提升几个百分点。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}