参考资料:《深度学习框架PyTorch:入门与实践》
注意:书籍作者表示要注意编程简洁,所以本文的网络搭建方式是可以借鉴的。
完整代码
部分讲解见注释:
from torch import nn
import torch as t
from torch.nn import functional as F
# 实现子module:Residual Block
class ResidualBlock(nn.Module):
def __init__(self, inchannel, outchannel, stride=1, shortcut=None):
super().__init__()
self.left = nn.Sequential(
# padding=1指的是上下左右各补一行,如果是padding=(1, 1)则是
# 上下各补一行、左右各补一行(先高后宽)
nn.Conv2d(inchannel, outchannel, 3, stride, 1, bias=False),
nn.BatchNorm2d(outchannel),
# inplace=True是指原地操作,直接修改上层传递下来的tensor,节省内存
nn.ReLU(inplace=True),
nn.Conv2d(outchannel, outchannel, 3, 1, 1, bias=False),
nn.BatchNorm2d(outchannel))
self.right = shortcut
def forward(self, x):
out = self.left(x)
residual = x if self.right is None else self.right(x)
out += residual
return F.relu(out)
class ResNet(nn.Module):
def __init__(self, num_classes=1000):
super().__init__()
self.pre = nn.Sequential(
nn.Conv2d(3, 64, 7, 2, 3, bias=False),
nn.BatchNorm2d(64),
nn.ReLU(inplace=True),
# kernel_size, stride, padding
nn.MaxPool2d(3, 2, 1))
self.layer1 = self._make_layer(64, 128, 3)
self.layer2 = self._make_layer(128, 256, 4, stride=2)
self.layer3 = self._make_layer(256, 512, 6, stride=2)
self.layer4 = self._make_layer(512, 512, 3, stride=2)
self.fc = nn.Linear(512, num_classes)
def _make_layer(self, inchannel, outchannel, block_num, stride=1):
shortcut = nn.Sequential(
nn.Conv2d(inchannel, outchannel, 1, stride, bias=False),
nn.BatchNorm2d(outchannel))
layers = []
# 第一个把inchannel变成outchannel
layers.append(ResidualBlock(inchannel, outchannel, stride, shortcut))
for i in range(1, block_num):
layers.append(ResidualBlock(outchannel, outchannel))
# list a=[1,2,3],那么*a就是指1 2 3
return nn.Sequential(*layers)
def forward(self, x):
x = self.pre(x)
x = self.layer1(x)
x = self.layer2(x)
x = self.layer3(x)
x = self.layer4(x)
x = F.avg_pool2d(x, 7)
# 变为[batch, ...]
x = x.view(x.size(0), -1)
return self.fc(x)
测试前向传播
model = ResNet()
# 模拟224×224的图片,通道为3。开头的1是batch_size
input = t.randn(1, 3, 224, 224)
o = model(input)
o.shape
输出结果:
torch.Size([1, 1000])
对应分类问题的1000个类别。