神经网络—Sequential的使用

本文通过搭建一个简单的神经网络来学习Sequential的使用

---

目录

一、搭建神经网络

二、Sequential的使用

---

一、搭建神经网络

代码如下：

import torch
from torch import nn
from torch.nn import Conv2d, MaxPool2d, Flatten, Linear

#搭建神经网络
class Test(nn.Module):
    def __init__(self):               #初始化
        super(Test, self).__init__()      #继承父类
        self.conv1 = Conv2d(in_channels=3, out_channels=32, kernel_size=5, padding=2)        #第一次卷积
        self.maxpool1 = MaxPool2d(2)                #第一次池化
        self.conv2 = Conv2d(32, 32, 5, padding=2)   #第二次卷积
        self.maxpool2 = MaxPool2d(2)                #第二次池化
        self.conv3 = Conv2d(32, 64, 5, padding=2)   #第三次卷积
        self.maxpool3 = MaxPool2d(2)                #地三次池化
        self.flatten = Flatten()                    #torch.nn.Flatten展平成一维
        self.linear1 = Linear(1024, 64)             #第一次线性
        self.linear2 = Linear(64, 10)               #第二次线性

    def forward(self, x):
        x = self.conv1(x)
        x = self.maxpool1(x)
        x = self.conv2(x)
        x = self.maxpool2(x)
        x = self.conv3(x)
        x = self.maxpool3(x)
        x = self.flatten(x)
        x = self.linear1(x)
        x = self.linear2(x)
        return x

test = Test()    #创建网络，初始化
print(test)

input = torch.ones((64, 3, 32, 32))
output = test(input)
print(output.shape)

输出结果：
Test(
  (conv1): Conv2d(3, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
  (maxpool1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv2d(32, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
  (maxpool2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv3): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
  (maxpool3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (flatten): Flatten(start_dim=1, end_dim=-1)
  (linear1): Linear(in_features=1024, out_features=64, bias=True)
  (linear2): Linear(in_features=64, out_features=10, bias=True)
)
torch.Size([64, 10])

二、Sequential的使用

代码如下：

import torch
from torch import nn
from torch.nn import Conv2d, MaxPool2d, Flatten, Linear, Sequential


#搭建神经网络
class Test(nn.Module):
    def __init__(self):               #初始化
        super(Test, self).__init__()      #继承父类
        self.sequential = Sequential(
            Conv2d(in_channels=3, out_channels=32, kernel_size=5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 32, 5, padding=2),
            MaxPool2d(2),
            Conv2d(32, 64, 5, padding=2),
            MaxPool2d(2),
            Flatten(),
            Linear(1024, 64),
            Linear(64, 10)
        )
    def forward(self, x):
        x = self.sequential(x)
        return x

test = Test()    #创建网络，初始化
print(test)

input = torch.ones((64, 3, 32, 32))
output = test(input)
print(output.shape)

输出结果：
Test(
  (sequential): Sequential(
    (0): Conv2d(3, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
    (1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (2): Conv2d(32, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
    (3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (4): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
    (5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (6): Flatten(start_dim=1, end_dim=-1)
    (7): Linear(in_features=1024, out_features=64, bias=True)
    (8): Linear(in_features=64, out_features=10, bias=True)
  )
)
torch.Size([64, 10])

不难看出，sequential的使用，可以使得代码更加简洁。

使用tensorboard进行图像的可视化

writer = SummaryWriter("logs")
writer.add_graph(test, input)        #.add_graph() 查看计算图
writer.close()

使用tensorboard进行图像的可视化，可以很清楚的了解网络中的每一步计算的输入和输出。