站长资源脚本专栏

pytorch forward两个参数实例

整理:jimmy2024/11/16浏览2
简介以channel Attention Block为例子class CAB(nn.Module):def __init__(self, in_channels, out_channels):super(CAB, self).__init__()self.global_pooling = nn.Adap

以channel Attention Block为例子

class CAB(nn.Module):
 
  def __init__(self, in_channels, out_channels):
    super(CAB, self).__init__()
    self.global_pooling = nn.AdaptiveAvgPool2d(output_size=1)
    self.conv1 = nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=1, padding=0)
    self.relu = nn.ReLU()
    self.conv2 = nn.Conv2d(out_channels, out_channels, kernel_size=1, stride=1, padding=0)
    self.sigmod = nn.Sigmoid()
 
  def forward(self, x):
    x1, x2 = x # high, low
    x = torch.cat([x1,x2],dim=1)
    x = self.global_pooling(x)
    x = self.conv1(x)
    x = self.relu(x)
    x = self.conv2(x)
    x = self.sigmod(x)
    x2 = x * x2
    res = x2 + x1
    return res

以上这篇pytorch forward两个参数实例就是小编分享给大家的全部内容了,希望能给大家一个参考,也希望大家多多支持。