使用conv 1*1 卷积来改变通道数极其代码

代码

import torch
import torch.nn as nn

# 定义 1x1 卷积
conv1x1 = nn.Conv2d(in_channels=128, out_channels=256, kernel_size=1)

# 输入特征图，假设其形状为 (batch_size, in_channels, height, width)
input_tensor = torch.randn(1, 128, 64, 64)

# 应用 1x1 卷积
output_tensor = conv1x1(input_tensor)

# 输出的形状将为 (batch_size, out_channels, height, width)
print(output_tensor.shape)  # 输出: torch.Size([1, 256, 64, 64])

参考其他博文

【pytorch】1x1卷积