MobileNet学习（二）——Pytorch搭建v2网络

最新推荐文章于 2024-04-07 16:35:32 发布

ccc_irene

最新推荐文章于 2024-04-07 16:35:32 发布

阅读量504

点赞数

分类专栏：机器学习

本文链接：https://blog.csdn.net/qq_41433316/article/details/98244421

版权

先放上模型结构：
在这里插入图片描述
对于stride=1或2的情况有不同的结构，stride=2时，不采用shortcut：

import torch
import torch.nn as nn
import torch.nn.functional as F

class Bottleneck(nn.Module):
    def __init__(self, in_channels, out_channels, multiple, stride):
        super(Bottleneck, self).__init__()
        self.stride = stride
        self.conv1 = nn.Conv2d(
            in_channels=in_channels,
            out_channels=in_channels*multiple,
            kernel_size=1,
            stride=1,
            padding=0
        )
        self.bn1 = nn.BatchNorm2d(in_channels*multiple)
        self.conv2 = nn.Conv2d(
            in_channels=in_channels*multiple,
            out_channels=in_channels*multiple,
            kernel_size=3,
            stride=stride,
            padding=1,
            groups=in_channels*multiple,
        )
        self.bn2 = nn.BatchNorm2d(in_channels*multiple)
        self.conv3 = nn.Conv2d(
            in_channels=in_channels*multiple