以字典查询的方式搭建网络结构并用到nn.Sequential(*a)

import torch.nn as nn
a=[]
cfgs = {
    'vgg11': [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],
    }
in_channels=3
for key,value in cfgs.items():
    print("\nkey:"+key)
    print("value"+str(value))
    for v in value:
        if v=="M":
           a+=[nn.MaxPool2d(kernel_size=2,stride=2)]
        else:
           conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding_mode=1)
           a += [conv2d]
           in_channels = v
b = nn.Sequential(*a)
print(b)

运行结果为

key:vgg11
value[64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M']
Sequential(
  (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1))
  (1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (2): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1))
  (3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (4): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1))
  (5): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1))
  (6): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (7): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1))
  (8): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1))
  (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (10): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1))
  (11): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1))
  (12): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)

Process finished with exit code 0