PyTorch torch.nn.AvgPool2d 函数
PyTorch torch.nn 参考手册torch.nn.AvgPool2d 是 PyTorch 中的二维平均池化模块。
它对输入的每个窗口计算平均值，常用于下采样和特征汇聚。
函数定义
torch.nn.AvgPool2d(kernel_size, stride=None, padding=0, ceil_mode=False, count_include_pad=True)
参数说明
kernel_size: 池化窗口大小stride: 步长，默认为 kernel_sizepadding: 填充大小ceil_mode: 是否使用 ceil 计算输出尺寸
使用示例
示例 1: 基本用法
实例
import torch
import torch.nn as nn
pool = nn.AvgPool2d(kernel_size=2, stride=2)
x = torch.randn(1, 1, 4, 4)
output = pool(x)
print("输入:n", x.squeeze().tolist())
print("n输出:n", output.squeeze().tolist())
print("形状:", x.shape, "->", output.shape)
import torch.nn as nn
pool = nn.AvgPool2d(kernel_size=2, stride=2)
x = torch.randn(1, 1, 4, 4)
output = pool(x)
print("输入:n", x.squeeze().tolist())
print("n输出:n", output.squeeze().tolist())
print("形状:", x.shape, "->", output.shape)
示例 2: 全局平均池化
实例
import torch
import torch.nn as nn
# 全局平均池化
gap = nn.AdaptiveAvgPool2d(1)
x = torch.randn(4, 64, 16, 16)
out = gap(x)
print("输入:", x.shape)
print("输出:", out.shape)
print("每个通道的平均值数量:", out.numel() // 64)
import torch.nn as nn
# 全局平均池化
gap = nn.AdaptiveAvgPool2d(1)
x = torch.randn(4, 64, 16, 16)
out = gap(x)
print("输入:", x.shape)
print("输出:", out.shape)
print("每个通道的平均值数量:", out.numel() // 64)
示例 3: 对比 MaxPool
实例
import torch
import torch.nn as nn
x = torch.tensor([[[
[1, 2, 3, 4],
[5, 6, 7, 8],
[9, 10, 11, 12],
[13, 14, 15, 16]
]]], dtype=torch.float32)
maxpool = nn.MaxPool2d(2, 2)
avgpool = nn.AvgPool2d(2, 2)
print("输入:n", x[0, 0])
print("nMaxPool:n", maxpool(x)[0, 0])
print("nAvgPool:n", avgpool(x)[0, 0])
import torch.nn as nn
x = torch.tensor([[[
[1, 2, 3, 4],
[5, 6, 7, 8],
[9, 10, 11, 12],
[13, 14, 15, 16]
]]], dtype=torch.float32)
maxpool = nn.MaxPool2d(2, 2)
avgpool = nn.AvgPool2d(2, 2)
print("输入:n", x[0, 0])
print("nMaxPool:n", maxpool(x)[0, 0])
print("nAvgPool:n", avgpool(x)[0, 0])
使用场景
- 特征汇聚: 减少空间维度
- 全局平均池化: 替代全连接层
- 平滑特征减少噪声