LeNet

In [ ]:

import torch
import torch.nn as nn

import torch import torch.nn as nn

In [ ]:

class LeNet(nn.Module):
    def __init__ (self, num_label=10):
        super(LeNet, self).__init__()
        self.net = nn.Sequential(
            # C1 (1*32*32) 到 (6*28*28)
            nn.Conv2d(in_channels=1,out_channels=6,kernel_size=5,stride=1,padding=0),
            nn.ReLU(),
            # S2 (6*28*28) 到 (6*14*14)
            nn.AvgPool2d(kernel_size=2,stride=2),
            # C3 (6*14*14) 到 (16*10*10)
            nn.Conv2d(in_channels=6, out_channels=16,kernel_size=5, stride=1,padding=0),
            nn.ReLU(),
            # S4 (16*10*10) 到 (16*5*5)
            nn.AvgPool2d(kernel_size=2, stride=2),
            # C5 (16*5*5) 到 (120*1*1)
            nn.Conv2d(in_channels=16,out_channels=120,kernel_size=5,stride=1,padding=0),
            nn.ReLU(),
            # 拉成向量
            nn.Flatten(),
            # F6 120 到 84
            nn.Linear(120, 84),
            nn.ReLU(),
            # OUTPUT 84 到 10
            nn.Linear(84, num_label)
        )

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

class LeNet(nn.Module): def __init__ (self, num_label=10): super(LeNet, self).__init__() self.net = nn.Sequential( # C1 (1*32*32) 到 (6*28*28) nn.Conv2d(in_channels=1,out_channels=6,kernel_size=5,stride=1,padding=0), nn.ReLU(), # S2 (6*28*28) 到 (6*14*14) nn.AvgPool2d(kernel_size=2,stride=2), # C3 (6*14*14) 到 (16*10*10) nn.Conv2d(in_channels=6, out_channels=16,kernel_size=5, stride=1,padding=0), nn.ReLU(), # S4 (16*10*10) 到 (16*5*5) nn.AvgPool2d(kernel_size=2, stride=2), # C5 (16*5*5) 到 (120*1*1) nn.Conv2d(in_channels=16,out_channels=120,kernel_size=5,stride=1,padding=0), nn.ReLU(), # 拉成向量 nn.Flatten(), # F6 120 到 84 nn.Linear(120, 84), nn.ReLU(), # OUTPUT 84 到 10 nn.Linear(84, num_label) ) def forward(self, x): x = self.net(x) return x