4.1 池化层原理
① 最大池化层有时也被称为下采样。
② dilation为空洞卷积,如下图所示。
③ Ceil_model为当超出区域时,只取最左上角的值。
④ 池化使得数据由5 * 5 变为3 * 3,甚至1 * 1的,这样导致计算的参数会大大减小。例如1080P的电影经过池化的转为720P的电影、或360P的电影后,同样的网速下,视频更为不卡。
4.2 池化层处理数据
import torch
from torch import nn
from torch.nn import MaxPool2dinput = torch.tensor([[1,2,0,3,1],[0,1,2,3,1],[1,2,1,0,0],[5,2,3,1,1],[2,1,0,1,1]], dtype = torch.float32)
input = torch.reshape(input,(-1,1,5,5))
print(input.shape)class Tudui(nn.Module):def __init__(self):super(Tudui, self).__init__()self.maxpool = MaxPool2d(kernel_size=3, ceil_mode=True)def forward(self, input):output = self.maxpool(input)return outputtudui = Tudui()
output = tudui(input)
print(output)
结果:
4.3 池化层处理图片
import torch
import torchvision
from torch import nn
from torch.nn import MaxPool2d
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriterdataset = torchvision.datasets.CIFAR10("./dataset",train=False,transform=torchvision.transforms.ToTensor(),download=True)
dataloader = DataLoader(dataset, batch_size=64)class Tudui(nn.Module):def __init__(self):super(Tudui, self).__init__()self.maxpool = MaxPool2d(kernel_size=3, ceil_mode=True)def forward(self, input):output = self.maxpool(input)return outputtudui = Tudui()
writer = SummaryWriter("logs")
step = 0for data in dataloader:imgs, targets = datawriter.add_images("input", imgs, step)output = tudui(imgs)writer.add_images("output", output, step)step = step + 1
操作:
① 在 Anaconda 终端里面,激活py3.6.3环境,再输入 tensorboard --logdir=C:\Users\wangy\Desktop\03CV\logs 命令,将网址赋值浏览器的网址栏,回车,即可查看tensorboard显示日志情况。
结果: