残差网络(ResNet)

残差块

In [1]:
import torch
from torch import nn
from torch.nn import functional as F
from d2l import torch as d2l

class Residual(nn.Module):  
    def __init__(self, input_channels, num_channels, use_1x1conv=False,
                 strides=1):
        super().__init__()
        self.conv1 = nn.Conv2d(input_channels, num_channels, kernel_size=3,
                               padding=1, stride=strides)
        self.conv2 = nn.Conv2d(num_channels, num_channels, kernel_size=3,
                               padding=1)
        if use_1x1conv:
            self.conv3 = nn.Conv2d(input_channels, num_channels,
                                   kernel_size=1, stride=strides)
        else:
            self.conv3 = None
        self.bn1 = nn.BatchNorm2d(num_channels)
        self.bn2 = nn.BatchNorm2d(num_channels)
        self.relu = nn.ReLU(inplace=True)

    def forward(self, X):
        Y = F.relu(self.bn1(self.conv1(X)))
        Y = self.bn2(self.conv2(Y))
        if self.conv3:
            X = self.conv3(X)
        Y += X
        return F.relu(Y)

输入和输出形状一致

In [2]:
blk = Residual(3, 3)
X = torch.rand(4, 3, 6, 6)
Y = blk(X)
Y.shape
Out[2]:
torch.Size([4, 3, 6, 6])

增加输出通道数的同时,减半输出的高和宽

In [3]:
blk = Residual(3, 6, use_1x1conv=True, strides=2)
blk(X).shape
Out[3]:
torch.Size([4, 6, 3, 3])

ResNet模型

In [7]:
b1 = nn.Sequential(nn.Conv2d(1, 64, kernel_size=7, stride=2, padding=3),
                   nn.BatchNorm2d(64), nn.ReLU(),
                   nn.MaxPool2d(kernel_size=3, stride=2, padding=1))

def resnet_block(input_channels, num_channels, num_residuals,
                 first_block=False):
    blk = []
    for i in range(num_residuals):
        if i == 0 and not first_block:
            blk.append(
                Residual(input_channels, num_channels, use_1x1conv=True,
                         strides=2))
        else:
            blk.append(Residual(num_channels, num_channels))
    return blk

b2 = nn.Sequential(*resnet_block(64, 64, 2, first_block=True))
b3 = nn.Sequential(*resnet_block(64, 128, 2))
b4 = nn.Sequential(*resnet_block(128, 256, 2))
b5 = nn.Sequential(*resnet_block(256, 512, 2))

net = nn.Sequential(b1, b2, b3, b4, b5, nn.AdaptiveAvgPool2d((1, 1)),
                    nn.Flatten(), nn.Linear(512, 10))

观察一下ResNet中不同模块的输入形状是如何变化的

In [8]:
X = torch.rand(size=(1, 1, 224, 224))
for layer in net:
    X = layer(X)
    print(layer.__class__.__name__, 'output shape:\t', X.shape)
Sequential output shape:	 torch.Size([1, 64, 56, 56])
Sequential output shape:	 torch.Size([1, 64, 56, 56])
Sequential output shape:	 torch.Size([1, 128, 28, 28])
Sequential output shape:	 torch.Size([1, 256, 14, 14])
Sequential output shape:	 torch.Size([1, 512, 7, 7])
AdaptiveAvgPool2d output shape:	 torch.Size([1, 512, 1, 1])
Flatten output shape:	 torch.Size([1, 512])
Linear output shape:	 torch.Size([1, 10])

训练模型

In [9]:
lr, num_epochs, batch_size = 0.05, 10, 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size, resize=96)
d2l.train_ch6(net, train_iter, test_iter, num_epochs, lr, d2l.try_gpu())
loss 0.023, train acc 0.993, test acc 0.912
4687.2 examples/sec on cuda:0
2021-07-09T05:34:07.695935 image/svg+xml Matplotlib v3.3.4, https://matplotlib.org/