pytorch实现特殊的Module--Sqeuential三种写法_Python

pytorch实现特殊的Module--Sqeuential三种写法

2020-04-21 09:40fly_Xiaoma Python

今天小编就为大家分享一篇pytorch实现特殊的Module--Sqeuential三种写法。具有很好的参考价值，希望对大家有所帮助。一起跟随小编过来看看吧

我就废话不多说了，直接上代码吧！

				?

									# -*- coding: utf-8 -*-

									#@Time  :2019/7/1 13:34

									#@Author :XiaoMa

									import torch as t

									from torch import nn

									#Sequential的三种写法

									net1=nn.Sequential()

									net1.add_module('conv',nn.Conv2d(3,3,3))  #Conv2D(输入通道数，输出通道数，卷积核大小)

									net1.add_module('batchnorm',nn.BatchNorm2d(3))  #BatchNorm2d(特征数)

									net1.add_module('activation_layer',nn.ReLU())

									net2=nn.Sequential(nn.Conv2d(3,3,3),

									          nn.BatchNorm2d(3),

									          nn.ReLU()

									          )

									from collections import OrderedDict

									net3=nn.Sequential(OrderedDict([

									  ('conv1',nn.Conv2d(3,3,3)),

									  ('bh1',nn.BatchNorm2d(3)),

									  ('al',nn.ReLU())

									]))

									print('net1',net1)

									print('net2',net2)

									print('net3',net3)

									#可根据名字或序号取出子module

									print(net1.conv,net2[0],net3.conv1)

输出结果：

				?

									net1 Sequential(

									 (conv): Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1))

									 (batchnorm): BatchNorm2d(3, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)

									 (activation_layer): ReLU()

									)

									net2 Sequential(

									 (0): Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1))

									 (1): BatchNorm2d(3, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)

									 (2): ReLU()

									)

									net3 Sequential(

									 (conv1): Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1))

									 (bh1): BatchNorm2d(3, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)

									 (al): ReLU()

									)

									Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1)) 

									Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1)) 

									Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1))