（八）神经网络-池化层

最新推荐文章于 2024-09-03 18:12:49 发布

千禧霓虹

最新推荐文章于 2024-09-03 18:12:49 发布

阅读量342

点赞数 7

分类专栏： PyTorch快速入门文章标签：神经网络人工智能深度学习

本文链接：https://blog.csdn.net/qq_46093326/article/details/135670171

版权

PyTorch快速入门专栏收录该内容

16 篇文章 0 订阅

订阅专栏

本文详细介绍了PyTorch中MaxPool2d和MaxPool3d类，包括其工作原理、参数设置以及在CIFAR10数据集上的实战应用，同时展示了如何使用TensorBoard可视化输入和处理后的图像。

摘要由CSDN通过智能技术生成

1. Pooling Layers下的常用类

nn.MaxPool2d

Applies a 2D max pooling over an input signal composed of several input planes.

nn.MaxPool3d

Applies a 3D max pooling over an input signal composed of several input planes.

nn.AvgPool2d

Applies a 2D average pooling over an input signal composed of several input planes.

nn.AvgPool3d

Applies a 3D average pooling over an input signal composed of several input planes.

2.最大池化

2.1 MaxPool2D类介绍

在这里插入图片描述

2.2 参数相关介绍

在这里插入图片描述

2.3 最大池化原理

stride = 3

2.4 MaxPool2D代码实战

from torch.utils.data import DataLoader
from torchvision import transforms, datasets
from torch.utils.tensorboard import SummaryWriter
from torch import nn


class My_module(nn.Module):
    def __init__(self, *args, **kwargs) -> None:
        super().__init__(*args, **kwargs)
        self.maxpool = nn.MaxPool2d(kernel_size=3, ceil_mode=True)

    def forward(self, x):
        output = self.maxpool(x)
        return output


my_module = My_module()
writer = SummaryWriter(log_dir="MaxPool")
test_dataset = datasets.CIFAR10(root="datasets", transform=transforms.ToTensor(), download=True)
test_dataloader = DataLoader(dataset=test_dataset, batch_size=64, shuffle=True, drop_last=False)
step = 0
for data in test_dataloader:
    imgs, labels = data
    imgs_maxpool = my_module(imgs)
    writer.add_images(tag="Before_MaxPool", img_tensor=imgs, global_step=step)
    writer.add_images(tag="After_MaxPool", img_tensor=imgs_maxpool, global_step=step)
writer.close()