小土堆Pytorch学习笔记（九、卷积操作和卷积层）

一、卷积操作

由以下输入图像和卷积核做卷积操作

代码如下：

import torch

input = 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]])

kernel = torch.tensor([[1, 2, 1],
                       [0, 1, 0],
                       [2, 1, 0]])

print(input.shape)
print(kernel.shape)

运行结果如下：

torch.Size([5, 5])
torch.Size([3, 3])

再由cov2d的参数规定：输入有四个参数，而运行结果显示只有两个（只有高和宽），所以得改成四维，采用torch.reshape( )

运行如下代码：

import torch
import torch.nn.functional as F
input = 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]])

kernel = torch.tensor([[1, 2, 1],
                       [0, 1, 0],
                       [2, 1, 0]])
input = torch.reshape(input, [1, 1, 5, 5])
kernel = torch.reshape(kernel, [1, 1, 3, 3])

print(input.shape)
print(kernel.shape)

output = F.conv2d(input, kernel, stride=1)
print(output)

结果如下：

torch.Size([1, 1, 5, 5])
torch.Size([1, 1, 3, 3])
tensor([[[[10, 12, 12],
          [18, 16, 16],
          [13,  9,  3]]]])

再设置padding参数使得输出图像和输入图像大小一样：

output = F.conv2d(input, kernel, stride=1)
print(output)

运行结果如下：

tensor([[[[ 1,  3,  4, 10,  8],
          [ 5, 10, 12, 12,  6],
          [ 7, 18, 16, 16,  8],
          [11, 13,  9,  3,  4],
          [14, 13,  9,  7,  4]]]])

完整代码：

import torch
import torch.nn.functional as F
input = 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]])

kernel = torch.tensor([[1, 2, 1],
                       [0, 1, 0],
                       [2, 1, 0]])
input = torch.reshape(input, [1, 1, 5, 5])
kernel = torch.reshape(kernel, [1, 1, 3, 3])

print(input.shape)
print(kernel.shape)

output = F.conv2d(input, kernel, stride=1)
print(output)

output2 = F.conv2d(input, kernel, stride=1, padding=1)
print(output2)

二、卷积层

打开Pytorch官网查看卷积层，由于输入图像是二维的，看Conv2d


再搭建一个神经网络：
准备dataset和dataloader，dataset采用Pytorch内置的数据集，跟前面的视频的内容一样，然后dataloader采用继承DataLoader，再定义一个神经网络，重写__init__( )方法和__forward__方法。代码如下：

import torch
import torchvision
from torch import nn
from torch.nn import Conv2d
from torch.utils.data import DataLoader

dataset = torchvision.datasets.CIFAR10("../data", train=False, transform=torchvision.transforms.ToTensor(), download=True)
dataloader = DataLoader(dataset, batch_size=64)


class Cow(nn.Module):
    def __init__(self):
        super(Cow, self).__init__()
        self.conv1 = Conv2d(in_channels=3, out_channels=6, kernel_size=3, stride=1, padding=0)

    def forward(self, x):
        x = self.conv1(x)
        return x


cow = Cow()
print(cow)

敲代码时，注意给transform设置参数时ToTensor后面要加（），不然会下载很久以及其它错误。运行结果如下;

Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to ../data\cifar-10-python.tar.gz
100%|██████████| 170498071/170498071 [07:00<00:00, 405616.21it/s]
Extracting ../data\cifar-10-python.tar.gz to ../data
Cow(
  (conv1): Conv2d(3, 6, kernel_size=(3, 3), stride=(1, 1))
)

再运行下面代码：

import torch
import torchvision
from torch import nn
from torch.nn import Conv2d
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

dataset = torchvision.datasets.CIFAR10("../data", train=False, transform=torchvision.transforms.ToTensor(), download=True)
dataloader = DataLoader(dataset, batch_size=64)


class Cow(nn.Module):
    def __init__(self):
        super(Cow, self).__init__()
        self.conv1 = Conv2d(in_channels=3, out_channels=6, kernel_size=3, stride=1, padding=0)

    def forward(self, x):
        x = self.conv1(x)
        return x


cow = Cow()
# print(cow)
writer = SummaryWriter("../logs2")
step = 0
for data in dataloader:
    imgs, targets = data
    output = cow(imgs)
    print(imgs.shape)
    print(output.shape)
    writer.add_images("input", imgs, step)
    output = torch.reshape(output, (-1, 3, 30, 30))
    writer.add_images("output", output, step)

    step = step + 1

打开终端，输入命令：tensorboard --logdir=“…/logs2”，注意是在根目录下创建的"logs2"，所以得用"…/"指令，表示返回到根目录。tensorboard结果如下：