J9学习打卡笔记

• 🍨 本文为🔗365天深度学习训练营 中的学习记录博客
• 🍖 原作者：K同学啊

import os, PIL, random, pathlib
import torch
import torch.nn as nn
import torchvision.transforms as transforms
import torchvision
from torchvision import transforms, datasets
import torch.nn.functional as Fdevice = torch.device("cuda" if torch.cuda.is_available() else "cpu")device

device(type='cuda')

网络结构

import torch.nn.functional as Fclass BasicConv2d(nn.Module):def __init__(self, in_channels, out_channels, **kwargs):super(BasicConv2d, self).__init__()self.conv = nn.Conv2d(in_channels, out_channels, bias=False, **kwargs)self.bn = nn.BatchNorm2d(out_channels, eps=0.001)def forward(self, x):x = self.conv(x)x = self.bn(x)return F.relu(x, inplace=True)class InceptionV3(nn.Module):def __init__(self, num_classes=1000, aux_logits=False, transform_input=False):super(InceptionV3, self).__init__()self.aux_logits = aux_logitsself.transform_input = transform_inputself.Conv2d_1a_3x3 = BasicConv2d(3, 32, kernel_size=3, stride=2)self.Conv2d_2a_3x3 = BasicConv2d(32, 32, kernel_size=3)self.Conv2d_2b_3x3 = BasicConv2d(32, 64, kernel_size=3, padding=1)self.Conv2d_3b_1x1 = BasicConv2d(64, 80, kernel_size=1)self.Conv2d_4a_3x3 = BasicConv2d(80, 192, kernel_size=3)self.Mixed_5b = InceptionA(192, pool_features=32)self.Mixed_5c = InceptionA(256, pool_features=64)self.Mixed_5d = InceptionA(288, pool_features=64)self.Mixed_6a = ReductionA(288)self.Mixed_6b = InceptionB(768, channels_7x7=128)self.Mixed_6c = InceptionB(768, channels_7x7=160)self.Mixed_6d = InceptionB(768, channels_7x7=160)self.Mixed_6e = InceptionB(768, channels_7x7=192)if aux_logits:self.AuxLogits = InceptionAux(768, num_classes)self.Mixed_7a = ReductionB(768)self.Mixed_7b = InceptionC(1280)self.Mixed_7c = InceptionC(2048)self.fc = nn.Linear(2048, num_classes)def forward(self, x):if self.transform_input: # 1x = x.clone()x[:, 0] = x[:, 0] * (0.229 / 0.5) + (0.485 - 0.5) / 0.5x[:, 1] = x[:, 1] * (0.224 / 0.5) + (0.456 - 0.5) / 0.5x[:, 2] = x[:, 2] * (0.225 / 0.5) + (0.406 - 0.5) / 0.5# 299 x 299 x 3x = self.Conv2d_1a_3x3(x)# 149 x 149 x 32x = self.Conv2d_2a_3x3(x)# 147 x 147 x 32x = self.Conv2d_2b_3x3(x)# 147 x 147 x 64x = F.max_pool2d(x, kernel_size=3, stride=2)# 73 x 73 x 64x = self.Conv2d_3b_1x1(x)# 73 x 73 x 80x = self.Conv2d_4a_3x3(x)# 71 x 71 x 192x = F.max_pool2d(x, kernel_size=3, stride=2)# 35 x 35 x 192x = self.Mixed_5b(x)# 35 x 35 x 256x = self.Mixed_5c(x)# 35 x 35 x 288x = self.Mixed_5d(x)# 35 x 35 x 288x = self.Mixed_6a(x)# 17 x 17 x 768x = self.Mixed_6b(x)# 17 x 17 x 768x = self.Mixed_6c(x)# 17 x 17 x 768x = self.Mixed_6d(x)# 17 x 17 x 768x = self.Mixed_6e(x)# 17 x 17 x 768if self.training and self.aux_logits:aux = self.AuxLogits(x)# 17 x 17 x 768x = self.Mixed_7a(x)# 8 x 8 x 1280x = self.Mixed_7b(x)# 8 x 8 x 2048x = self.Mixed_7c(x)# 8 x 8 x 2048x = F.avg_pool2d(x, kernel_size=8)# 1 x 1 x 2048x = F.dropout(x, training=self.training)# 1 x 1 x 2048x = x.view(x.size(0), -1)# 2048x = self.fc(x)# 1000 (num_classes)if self.training and self.aux_logits:return x, auxreturn x

InceptionA

class InceptionA(nn.Module):def __init__(self, in_channels, pool_features):super(InceptionA, self).__init__()self.branch1x1 = BasicConv2d(in_channels, 64, kernel_size=1) # 1self.branch5x5_1 = BasicConv2d(in_channels, 48, kernel_size=1)self.branch5x5_2 = BasicConv2d(48, 64, kernel_size=5, padding=2)self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, padding=1)self.branch_pool = BasicConv2d(in_channels, pool_features, kernel_size=1)def forward(self, x):branch1x1 = self.branch1x1(x)branch5x5 = self.branch5x5_1(x)branch5x5 = self.branch5x5_2(branch5x5)branch3x3dbl = self.branch3x3dbl_1(x)branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)branch_pool = self.branch_pool(branch_pool)outputs = [branch1x1, branch5x5, branch3x3dbl, branch_pool]return torch.cat(outputs, 1)

InceptionB

class InceptionB(nn.Module):def __init__(self, in_channels, channels_7x7):super(InceptionB, self).__init__()self.branch1x1 = BasicConv2d(in_channels, 192, kernel_size=1)c7 = channels_7x7self.branch7x7_1 = BasicConv2d(in_channels, c7, kernel_size=1)self.branch7x7_2 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))self.branch7x7_3 = BasicConv2d(c7, 192, kernel_size=(7, 1), padding=(3, 0))self.branch7x7dbl_1 = BasicConv2d(in_channels, c7, kernel_size=1)self.branch7x7dbl_2 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))self.branch7x7dbl_3 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))self.branch7x7dbl_4 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))self.branch7x7dbl_5 = BasicConv2d(c7, 192, kernel_size=(1, 7), padding=(0, 3))self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)def forward(self, x):branch1x1 = self.branch1x1(x)branch7x7 = self.branch7x7_1(x)branch7x7 = self.branch7x7_2(branch7x7)branch7x7 = self.branch7x7_3(branch7x7)branch7x7dbl = self.branch7x7dbl_1(x)branch7x7dbl = self.branch7x7dbl_2(branch7x7dbl)branch7x7dbl = self.branch7x7dbl_3(branch7x7dbl)branch7x7dbl = self.branch7x7dbl_4(branch7x7dbl)branch7x7dbl = self.branch7x7dbl_5(branch7x7dbl)branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)branch_pool = self.branch_pool(branch_pool)outputs = [branch1x1, branch7x7, branch7x7dbl, branch_pool]return torch.cat(outputs, 1)

InceptionC

class InceptionC(nn.Module):def __init__(self, in_channels):super(InceptionC, self).__init__()self.branch1x1 = BasicConv2d(in_channels, 320, kernel_size=1)self.branch3x3_1 = BasicConv2d(in_channels, 384, kernel_size=1)self.branch3x3_2a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))self.branch3x3_2b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))self.branch3x3dbl_1 = BasicConv2d(in_channels, 448, kernel_size=1)self.branch3x3dbl_2 = BasicConv2d(448, 384, kernel_size=3, padding=1)self.branch3x3dbl_3a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))self.branch3x3dbl_3b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)def forward(self, x):branch1x1 = self.branch1x1(x)branch3x3 = self.branch3x3_1(x)branch3x3 = [self.branch3x3_2a(branch3x3),self.branch3x3_2b(branch3x3),]branch3x3 = torch.cat(branch3x3, 1)branch3x3dbl = self.branch3x3dbl_1(x)branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)branch3x3dbl = [self.branch3x3dbl_3a(branch3x3dbl),self.branch3x3dbl_3b(branch3x3dbl),]branch3x3dbl = torch.cat(branch3x3dbl, 1)branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)branch_pool = self.branch_pool(branch_pool)outputs = [branch1x1, branch3x3, branch3x3dbl, branch_pool]return torch.cat(outputs, 1)

ReductionA

class ReductionA(nn.Module):def __init__(self, in_channels):super(ReductionA, self).__init__()self.branch3x3 = BasicConv2d(in_channels, 384, kernel_size=3, stride=2)self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, stride=2)def forward(self, x):branch3x3 = self.branch3x3(x)branch3x3dbl = self.branch3x3dbl_1(x)branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)outputs = [branch3x3, branch3x3dbl, branch_pool]return torch.cat(outputs, 1)

ReductionB

class ReductionB(nn.Module):def __init__(self, in_channels):super(ReductionB, self).__init__()self.branch3x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)self.branch3x3_2 = BasicConv2d(192, 320, kernel_size=3, stride=2)self.branch7x7x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)self.branch7x7x3_2 = BasicConv2d(192, 192, kernel_size=(1, 7), padding=(0, 3))self.branch7x7x3_3 = BasicConv2d(192, 192, kernel_size=(7, 1), padding=(3, 0))self.branch7x7x3_4 = BasicConv2d(192, 192, kernel_size=3, stride=2)def forward(self, x):branch3x3 = self.branch3x3_1(x)branch3x3 = self.branch3x3_2(branch3x3)branch7x7x3 = self.branch7x7x3_1(x)branch7x7x3 = self.branch7x7x3_2(branch7x7x3)branch7x7x3 = self.branch7x7x3_3(branch7x7x3)branch7x7x3 = self.branch7x7x3_4(branch7x7x3)branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)outputs = [branch3x3, branch7x7x3, branch_pool]return torch.cat(outputs, 1)

辅助分支

class InceptionAux(nn.Module):def __init__(self, in_channels, num_classes):super(InceptionAux, self).__init__()self.conv0 = BasicConv2d(in_channels, 128, kernel_size=1)self.conv1 = BasicConv2d(128, 768, kernel_size=5)self.conv1.stddev = 0.01self.fc = nn.Linear(768, num_classes)self.fc.stddev = 0.001def forward(self, x):# 17 x 17 x 768x = F.avg_pool2d(x, kernel_size=5, stride=3)# 5 x 5 x 768x = self.conv0(x)# 5 x 5 x 128x = self.conv1(x)# 1 x 1 x 768x = x.view(x.size(0), -1)# 768x = self.fc(x)# 1000return x

device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))model = InceptionV3(4).to(device)
model

Using cuda deviceInceptionV3((Conv2d_1a_3x3): BasicConv2d((conv): Conv2d(3, 32, kernel_size=(3, 3), stride=(2, 2), bias=False)(bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(Conv2d_2a_3x3): BasicConv2d((conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), bias=False)(bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(Conv2d_2b_3x3): BasicConv2d((conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(Conv2d_3b_1x1): BasicConv2d((conv): Conv2d(64, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(80, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(Conv2d_4a_3x3): BasicConv2d((conv): Conv2d(80, 192, kernel_size=(3, 3), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(Mixed_5b): InceptionA((branch1x1): BasicConv2d((conv): Conv2d(192, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch5x5_1): BasicConv2d((conv): Conv2d(192, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch5x5_2): BasicConv2d((conv): Conv2d(48, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_1): BasicConv2d((conv): Conv2d(192, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_2): BasicConv2d((conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3): BasicConv2d((conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_5c): InceptionA((branch1x1): BasicConv2d((conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch5x5_1): BasicConv2d((conv): Conv2d(256, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch5x5_2): BasicConv2d((conv): Conv2d(48, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_1): BasicConv2d((conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_2): BasicConv2d((conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3): BasicConv2d((conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_5d): InceptionA((branch1x1): BasicConv2d((conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch5x5_1): BasicConv2d((conv): Conv2d(288, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch5x5_2): BasicConv2d((conv): Conv2d(48, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_1): BasicConv2d((conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_2): BasicConv2d((conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3): BasicConv2d((conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_6a): ReductionA((branch3x3): BasicConv2d((conv): Conv2d(288, 384, kernel_size=(3, 3), stride=(2, 2), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_1): BasicConv2d((conv): Conv2d(288, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_2): BasicConv2d((conv): Conv2d(64, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3): BasicConv2d((conv): Conv2d(96, 96, kernel_size=(3, 3), stride=(2, 2), bias=False)(bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_6b): InceptionB((branch1x1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_1): BasicConv2d((conv): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_2): BasicConv2d((conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_3): BasicConv2d((conv): Conv2d(128, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_1): BasicConv2d((conv): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_2): BasicConv2d((conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_3): BasicConv2d((conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_4): BasicConv2d((conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_5): BasicConv2d((conv): Conv2d(128, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_6c): InceptionB((branch1x1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_1): BasicConv2d((conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_2): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_3): BasicConv2d((conv): Conv2d(160, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_1): BasicConv2d((conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_2): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_3): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_4): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_5): BasicConv2d((conv): Conv2d(160, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_6d): InceptionB((branch1x1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_1): BasicConv2d((conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_2): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_3): BasicConv2d((conv): Conv2d(160, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_1): BasicConv2d((conv): Conv2d(768, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_2): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_3): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_4): BasicConv2d((conv): Conv2d(160, 160, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_5): BasicConv2d((conv): Conv2d(160, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_6e): InceptionB((branch1x1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_2): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7_3): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_2): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_3): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_4): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7dbl_5): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_7a): ReductionB((branch3x3_1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_2): BasicConv2d((conv): Conv2d(192, 320, kernel_size=(3, 3), stride=(2, 2), bias=False)(bn): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7x3_1): BasicConv2d((conv): Conv2d(768, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7x3_2): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7x3_3): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch7x7x3_4): BasicConv2d((conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(2, 2), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_7b): InceptionC((branch1x1): BasicConv2d((conv): Conv2d(1280, 320, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_1): BasicConv2d((conv): Conv2d(1280, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_2a): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_2b): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_1): BasicConv2d((conv): Conv2d(1280, 448, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(448, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_2): BasicConv2d((conv): Conv2d(448, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3a): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3b): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(1280, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(Mixed_7c): InceptionC((branch1x1): BasicConv2d((conv): Conv2d(2048, 320, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_1): BasicConv2d((conv): Conv2d(2048, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_2a): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3_2b): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_1): BasicConv2d((conv): Conv2d(2048, 448, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(448, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_2): BasicConv2d((conv): Conv2d(448, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3a): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch3x3dbl_3b): BasicConv2d((conv): Conv2d(384, 384, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)(bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True))(branch_pool): BasicConv2d((conv): Conv2d(2048, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)))(fc): Linear(in_features=2048, out_features=4, bias=True)
)

# 统计模型参数量以及其他指标
import torchsummary as summary
summary.summary(model, (3, 299, 299))

----------------------------------------------------------------Layer (type)               Output Shape         Param #
================================================================Conv2d-1         [-1, 32, 149, 149]             864BatchNorm2d-2         [-1, 32, 149, 149]              64BasicConv2d-3         [-1, 32, 149, 149]               0Conv2d-4         [-1, 32, 147, 147]           9,216BatchNorm2d-5         [-1, 32, 147, 147]              64BasicConv2d-6         [-1, 32, 147, 147]               0Conv2d-7         [-1, 64, 147, 147]          18,432BatchNorm2d-8         [-1, 64, 147, 147]             128BasicConv2d-9         [-1, 64, 147, 147]               0Conv2d-10           [-1, 80, 73, 73]           5,120BatchNorm2d-11           [-1, 80, 73, 73]             160BasicConv2d-12           [-1, 80, 73, 73]               0Conv2d-13          [-1, 192, 71, 71]         138,240BatchNorm2d-14          [-1, 192, 71, 71]             384BasicConv2d-15          [-1, 192, 71, 71]               0Conv2d-16           [-1, 64, 35, 35]          12,288BatchNorm2d-17           [-1, 64, 35, 35]             128BasicConv2d-18           [-1, 64, 35, 35]               0Conv2d-19           [-1, 48, 35, 35]           9,216BatchNorm2d-20           [-1, 48, 35, 35]              96BasicConv2d-21           [-1, 48, 35, 35]               0Conv2d-22           [-1, 64, 35, 35]          76,800BatchNorm2d-23           [-1, 64, 35, 35]             128BasicConv2d-24           [-1, 64, 35, 35]               0Conv2d-25           [-1, 64, 35, 35]          12,288BatchNorm2d-26           [-1, 64, 35, 35]             128BasicConv2d-27           [-1, 64, 35, 35]               0Conv2d-28           [-1, 96, 35, 35]          55,296BatchNorm2d-29           [-1, 96, 35, 35]             192BasicConv2d-30           [-1, 96, 35, 35]               0Conv2d-31           [-1, 96, 35, 35]          82,944BatchNorm2d-32           [-1, 96, 35, 35]             192BasicConv2d-33           [-1, 96, 35, 35]               0Conv2d-34           [-1, 32, 35, 35]           6,144BatchNorm2d-35           [-1, 32, 35, 35]              64BasicConv2d-36           [-1, 32, 35, 35]               0InceptionA-37          [-1, 256, 35, 35]               0Conv2d-38           [-1, 64, 35, 35]          16,384BatchNorm2d-39           [-1, 64, 35, 35]             128BasicConv2d-40           [-1, 64, 35, 35]               0Conv2d-41           [-1, 48, 35, 35]          12,288BatchNorm2d-42           [-1, 48, 35, 35]              96BasicConv2d-43           [-1, 48, 35, 35]               0Conv2d-44           [-1, 64, 35, 35]          76,800BatchNorm2d-45           [-1, 64, 35, 35]             128BasicConv2d-46           [-1, 64, 35, 35]               0Conv2d-47           [-1, 64, 35, 35]          16,384BatchNorm2d-48           [-1, 64, 35, 35]             128BasicConv2d-49           [-1, 64, 35, 35]               0Conv2d-50           [-1, 96, 35, 35]          55,296BatchNorm2d-51           [-1, 96, 35, 35]             192BasicConv2d-52           [-1, 96, 35, 35]               0Conv2d-53           [-1, 96, 35, 35]          82,944BatchNorm2d-54           [-1, 96, 35, 35]             192BasicConv2d-55           [-1, 96, 35, 35]               0Conv2d-56           [-1, 64, 35, 35]          16,384BatchNorm2d-57           [-1, 64, 35, 35]             128BasicConv2d-58           [-1, 64, 35, 35]               0InceptionA-59          [-1, 288, 35, 35]               0Conv2d-60           [-1, 64, 35, 35]          18,432BatchNorm2d-61           [-1, 64, 35, 35]             128BasicConv2d-62           [-1, 64, 35, 35]               0Conv2d-63           [-1, 48, 35, 35]          13,824BatchNorm2d-64           [-1, 48, 35, 35]              96BasicConv2d-65           [-1, 48, 35, 35]               0Conv2d-66           [-1, 64, 35, 35]          76,800BatchNorm2d-67           [-1, 64, 35, 35]             128BasicConv2d-68           [-1, 64, 35, 35]               0Conv2d-69           [-1, 64, 35, 35]          18,432BatchNorm2d-70           [-1, 64, 35, 35]             128BasicConv2d-71           [-1, 64, 35, 35]               0Conv2d-72           [-1, 96, 35, 35]          55,296BatchNorm2d-73           [-1, 96, 35, 35]             192BasicConv2d-74           [-1, 96, 35, 35]               0Conv2d-75           [-1, 96, 35, 35]          82,944BatchNorm2d-76           [-1, 96, 35, 35]             192BasicConv2d-77           [-1, 96, 35, 35]               0Conv2d-78           [-1, 64, 35, 35]          18,432BatchNorm2d-79           [-1, 64, 35, 35]             128BasicConv2d-80           [-1, 64, 35, 35]               0InceptionA-81          [-1, 288, 35, 35]               0Conv2d-82          [-1, 384, 17, 17]         995,328BatchNorm2d-83          [-1, 384, 17, 17]             768BasicConv2d-84          [-1, 384, 17, 17]               0Conv2d-85           [-1, 64, 35, 35]          18,432BatchNorm2d-86           [-1, 64, 35, 35]             128BasicConv2d-87           [-1, 64, 35, 35]               0Conv2d-88           [-1, 96, 35, 35]          55,296BatchNorm2d-89           [-1, 96, 35, 35]             192BasicConv2d-90           [-1, 96, 35, 35]               0Conv2d-91           [-1, 96, 17, 17]          82,944BatchNorm2d-92           [-1, 96, 17, 17]             192BasicConv2d-93           [-1, 96, 17, 17]               0ReductionA-94          [-1, 768, 17, 17]               0Conv2d-95          [-1, 192, 17, 17]         147,456BatchNorm2d-96          [-1, 192, 17, 17]             384BasicConv2d-97          [-1, 192, 17, 17]               0Conv2d-98          [-1, 128, 17, 17]          98,304BatchNorm2d-99          [-1, 128, 17, 17]             256BasicConv2d-100          [-1, 128, 17, 17]               0Conv2d-101          [-1, 128, 17, 17]         114,688BatchNorm2d-102          [-1, 128, 17, 17]             256BasicConv2d-103          [-1, 128, 17, 17]               0Conv2d-104          [-1, 192, 17, 17]         172,032BatchNorm2d-105          [-1, 192, 17, 17]             384BasicConv2d-106          [-1, 192, 17, 17]               0Conv2d-107          [-1, 128, 17, 17]          98,304BatchNorm2d-108          [-1, 128, 17, 17]             256BasicConv2d-109          [-1, 128, 17, 17]               0Conv2d-110          [-1, 128, 17, 17]         114,688BatchNorm2d-111          [-1, 128, 17, 17]             256BasicConv2d-112          [-1, 128, 17, 17]               0Conv2d-113          [-1, 128, 17, 17]         114,688BatchNorm2d-114          [-1, 128, 17, 17]             256BasicConv2d-115          [-1, 128, 17, 17]               0Conv2d-116          [-1, 128, 17, 17]         114,688BatchNorm2d-117          [-1, 128, 17, 17]             256BasicConv2d-118          [-1, 128, 17, 17]               0Conv2d-119          [-1, 192, 17, 17]         172,032BatchNorm2d-120          [-1, 192, 17, 17]             384BasicConv2d-121          [-1, 192, 17, 17]               0Conv2d-122          [-1, 192, 17, 17]         147,456BatchNorm2d-123          [-1, 192, 17, 17]             384BasicConv2d-124          [-1, 192, 17, 17]               0InceptionB-125          [-1, 768, 17, 17]               0Conv2d-126          [-1, 192, 17, 17]         147,456BatchNorm2d-127          [-1, 192, 17, 17]             384BasicConv2d-128          [-1, 192, 17, 17]               0Conv2d-129          [-1, 160, 17, 17]         122,880BatchNorm2d-130          [-1, 160, 17, 17]             320BasicConv2d-131          [-1, 160, 17, 17]               0Conv2d-132          [-1, 160, 17, 17]         179,200BatchNorm2d-133          [-1, 160, 17, 17]             320BasicConv2d-134          [-1, 160, 17, 17]               0Conv2d-135          [-1, 192, 17, 17]         215,040BatchNorm2d-136          [-1, 192, 17, 17]             384BasicConv2d-137          [-1, 192, 17, 17]               0Conv2d-138          [-1, 160, 17, 17]         122,880BatchNorm2d-139          [-1, 160, 17, 17]             320BasicConv2d-140          [-1, 160, 17, 17]               0Conv2d-141          [-1, 160, 17, 17]         179,200BatchNorm2d-142          [-1, 160, 17, 17]             320BasicConv2d-143          [-1, 160, 17, 17]               0Conv2d-144          [-1, 160, 17, 17]         179,200BatchNorm2d-145          [-1, 160, 17, 17]             320BasicConv2d-146          [-1, 160, 17, 17]               0Conv2d-147          [-1, 160, 17, 17]         179,200BatchNorm2d-148          [-1, 160, 17, 17]             320BasicConv2d-149          [-1, 160, 17, 17]               0Conv2d-150          [-1, 192, 17, 17]         215,040BatchNorm2d-151          [-1, 192, 17, 17]             384BasicConv2d-152          [-1, 192, 17, 17]               0Conv2d-153          [-1, 192, 17, 17]         147,456BatchNorm2d-154          [-1, 192, 17, 17]             384BasicConv2d-155          [-1, 192, 17, 17]               0InceptionB-156          [-1, 768, 17, 17]               0Conv2d-157          [-1, 192, 17, 17]         147,456BatchNorm2d-158          [-1, 192, 17, 17]             384BasicConv2d-159          [-1, 192, 17, 17]               0Conv2d-160          [-1, 160, 17, 17]         122,880BatchNorm2d-161          [-1, 160, 17, 17]             320BasicConv2d-162          [-1, 160, 17, 17]               0Conv2d-163          [-1, 160, 17, 17]         179,200BatchNorm2d-164          [-1, 160, 17, 17]             320BasicConv2d-165          [-1, 160, 17, 17]               0Conv2d-166          [-1, 192, 17, 17]         215,040BatchNorm2d-167          [-1, 192, 17, 17]             384BasicConv2d-168          [-1, 192, 17, 17]               0Conv2d-169          [-1, 160, 17, 17]         122,880BatchNorm2d-170          [-1, 160, 17, 17]             320BasicConv2d-171          [-1, 160, 17, 17]               0Conv2d-172          [-1, 160, 17, 17]         179,200BatchNorm2d-173          [-1, 160, 17, 17]             320BasicConv2d-174          [-1, 160, 17, 17]               0Conv2d-175          [-1, 160, 17, 17]         179,200BatchNorm2d-176          [-1, 160, 17, 17]             320BasicConv2d-177          [-1, 160, 17, 17]               0Conv2d-178          [-1, 160, 17, 17]         179,200BatchNorm2d-179          [-1, 160, 17, 17]             320BasicConv2d-180          [-1, 160, 17, 17]               0Conv2d-181          [-1, 192, 17, 17]         215,040BatchNorm2d-182          [-1, 192, 17, 17]             384BasicConv2d-183          [-1, 192, 17, 17]               0Conv2d-184          [-1, 192, 17, 17]         147,456BatchNorm2d-185          [-1, 192, 17, 17]             384BasicConv2d-186          [-1, 192, 17, 17]               0InceptionB-187          [-1, 768, 17, 17]               0Conv2d-188          [-1, 192, 17, 17]         147,456BatchNorm2d-189          [-1, 192, 17, 17]             384BasicConv2d-190          [-1, 192, 17, 17]               0Conv2d-191          [-1, 192, 17, 17]         147,456BatchNorm2d-192          [-1, 192, 17, 17]             384BasicConv2d-193          [-1, 192, 17, 17]               0Conv2d-194          [-1, 192, 17, 17]         258,048BatchNorm2d-195          [-1, 192, 17, 17]             384BasicConv2d-196          [-1, 192, 17, 17]               0Conv2d-197          [-1, 192, 17, 17]         258,048BatchNorm2d-198          [-1, 192, 17, 17]             384BasicConv2d-199          [-1, 192, 17, 17]               0Conv2d-200          [-1, 192, 17, 17]         147,456BatchNorm2d-201          [-1, 192, 17, 17]             384BasicConv2d-202          [-1, 192, 17, 17]               0Conv2d-203          [-1, 192, 17, 17]         258,048BatchNorm2d-204          [-1, 192, 17, 17]             384BasicConv2d-205          [-1, 192, 17, 17]               0Conv2d-206          [-1, 192, 17, 17]         258,048BatchNorm2d-207          [-1, 192, 17, 17]             384BasicConv2d-208          [-1, 192, 17, 17]               0Conv2d-209          [-1, 192, 17, 17]         258,048BatchNorm2d-210          [-1, 192, 17, 17]             384BasicConv2d-211          [-1, 192, 17, 17]               0Conv2d-212          [-1, 192, 17, 17]         258,048BatchNorm2d-213          [-1, 192, 17, 17]             384BasicConv2d-214          [-1, 192, 17, 17]               0Conv2d-215          [-1, 192, 17, 17]         147,456BatchNorm2d-216          [-1, 192, 17, 17]             384BasicConv2d-217          [-1, 192, 17, 17]               0InceptionB-218          [-1, 768, 17, 17]               0Conv2d-219          [-1, 192, 17, 17]         147,456BatchNorm2d-220          [-1, 192, 17, 17]             384BasicConv2d-221          [-1, 192, 17, 17]               0Conv2d-222            [-1, 320, 8, 8]         552,960BatchNorm2d-223            [-1, 320, 8, 8]             640BasicConv2d-224            [-1, 320, 8, 8]               0Conv2d-225          [-1, 192, 17, 17]         147,456BatchNorm2d-226          [-1, 192, 17, 17]             384BasicConv2d-227          [-1, 192, 17, 17]               0Conv2d-228          [-1, 192, 17, 17]         258,048BatchNorm2d-229          [-1, 192, 17, 17]             384BasicConv2d-230          [-1, 192, 17, 17]               0Conv2d-231          [-1, 192, 17, 17]         258,048BatchNorm2d-232          [-1, 192, 17, 17]             384BasicConv2d-233          [-1, 192, 17, 17]               0Conv2d-234            [-1, 192, 8, 8]         331,776BatchNorm2d-235            [-1, 192, 8, 8]             384BasicConv2d-236            [-1, 192, 8, 8]               0ReductionB-237           [-1, 1280, 8, 8]               0Conv2d-238            [-1, 320, 8, 8]         409,600BatchNorm2d-239            [-1, 320, 8, 8]             640BasicConv2d-240            [-1, 320, 8, 8]               0Conv2d-241            [-1, 384, 8, 8]         491,520BatchNorm2d-242            [-1, 384, 8, 8]             768BasicConv2d-243            [-1, 384, 8, 8]               0Conv2d-244            [-1, 384, 8, 8]         442,368BatchNorm2d-245            [-1, 384, 8, 8]             768BasicConv2d-246            [-1, 384, 8, 8]               0Conv2d-247            [-1, 384, 8, 8]         442,368BatchNorm2d-248            [-1, 384, 8, 8]             768BasicConv2d-249            [-1, 384, 8, 8]               0Conv2d-250            [-1, 448, 8, 8]         573,440BatchNorm2d-251            [-1, 448, 8, 8]             896BasicConv2d-252            [-1, 448, 8, 8]               0Conv2d-253            [-1, 384, 8, 8]       1,548,288BatchNorm2d-254            [-1, 384, 8, 8]             768BasicConv2d-255            [-1, 384, 8, 8]               0Conv2d-256            [-1, 384, 8, 8]         442,368BatchNorm2d-257            [-1, 384, 8, 8]             768BasicConv2d-258            [-1, 384, 8, 8]               0Conv2d-259            [-1, 384, 8, 8]         442,368BatchNorm2d-260            [-1, 384, 8, 8]             768BasicConv2d-261            [-1, 384, 8, 8]               0Conv2d-262            [-1, 192, 8, 8]         245,760BatchNorm2d-263            [-1, 192, 8, 8]             384BasicConv2d-264            [-1, 192, 8, 8]               0InceptionC-265           [-1, 2048, 8, 8]               0Conv2d-266            [-1, 320, 8, 8]         655,360BatchNorm2d-267            [-1, 320, 8, 8]             640BasicConv2d-268            [-1, 320, 8, 8]               0Conv2d-269            [-1, 384, 8, 8]         786,432BatchNorm2d-270            [-1, 384, 8, 8]             768BasicConv2d-271            [-1, 384, 8, 8]               0Conv2d-272            [-1, 384, 8, 8]         442,368BatchNorm2d-273            [-1, 384, 8, 8]             768BasicConv2d-274            [-1, 384, 8, 8]               0Conv2d-275            [-1, 384, 8, 8]         442,368BatchNorm2d-276            [-1, 384, 8, 8]             768BasicConv2d-277            [-1, 384, 8, 8]               0Conv2d-278            [-1, 448, 8, 8]         917,504BatchNorm2d-279            [-1, 448, 8, 8]             896BasicConv2d-280            [-1, 448, 8, 8]               0Conv2d-281            [-1, 384, 8, 8]       1,548,288BatchNorm2d-282            [-1, 384, 8, 8]             768BasicConv2d-283            [-1, 384, 8, 8]               0Conv2d-284            [-1, 384, 8, 8]         442,368BatchNorm2d-285            [-1, 384, 8, 8]             768BasicConv2d-286            [-1, 384, 8, 8]               0Conv2d-287            [-1, 384, 8, 8]         442,368BatchNorm2d-288            [-1, 384, 8, 8]             768BasicConv2d-289            [-1, 384, 8, 8]               0Conv2d-290            [-1, 192, 8, 8]         393,216BatchNorm2d-291            [-1, 192, 8, 8]             384BasicConv2d-292            [-1, 192, 8, 8]               0InceptionC-293           [-1, 2048, 8, 8]               0Linear-294                    [-1, 4]           8,196
================================================================
Total params: 21,793,764
Trainable params: 21,793,764
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 1.02
Forward/backward pass size (MB): 224.12
Params size (MB): 83.14
Estimated Total Size (MB): 308.28
----------------------------------------------------------------

# 导入数据
data_dir = r'C:\Users\11054\Desktop\kLearning\p1_data'
data_dir = pathlib.Path(data_dir)data_paths = list(data_dir.glob('*'))
classeNames = [str(path).split("\\")[-1] for path in data_paths]
print(classeNames)image_count = len(list(data_dir.glob('*/*')))
print("图片总数为：", image_count)

['cloudy', 'rain', 'shine', 'sunrise']
图片总数为： 1125

# 数据预处理
train_transforms = transforms.Compose([transforms.Resize([299, 299]),  # 将输入图片resize成统一尺寸# transforms.RandomHorizontalFlip(), # 随机水平翻转transforms.ToTensor(),  # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间transforms.Normalize(  # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛mean=[0.485, 0.456, 0.406],std=[0.229, 0.224, 0.225])  # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])test_transform = transforms.Compose([transforms.Resize([299, 299]),  # 将输入图片resize成统一尺寸transforms.ToTensor(),  # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间transforms.Normalize(  # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛mean=[0.485, 0.456, 0.406],std=[0.229, 0.224, 0.225])  # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])total_data = datasets.ImageFolder(data_dir, transform=train_transforms)
print(total_data.class_to_idx)# 划分数据集
train_size = int(0.8 * len(total_data))
test_size = len(total_data) - train_size
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])batch_size = 8  #根据自己的显卡，选择合适的batch_size大小
train_dl = torch.utils.data.DataLoader(train_dataset,batch_size=batch_size,shuffle=True,num_workers=0)
test_dl = torch.utils.data.DataLoader(test_dataset,batch_size=batch_size,shuffle=True,num_workers=0)
for X, y in test_dl:print("Shape of X [N, C, H, W]: ", X.shape)print("Shape of y: ", y.shape, y.dtype)break

{'cloudy': 0, 'rain': 1, 'shine': 2, 'sunrise': 3}
Shape of X [N, C, H, W]:  torch.Size([8, 3, 299, 299])
Shape of y:  torch.Size([8]) torch.int64

# 训练和测试函数
def train(dataloader, model, loss_fn, optimizer):size = len(dataloader.dataset)  # 训练集的大小num_batches = len(dataloader)  # 批次数目, (size/batch_size，向上取整)train_loss, train_acc = 0, 0  # 初始化训练损失和正确率for X, y in dataloader:  # 获取图片及其标签X, y = X.to(device), y.to(device)# 计算预测误差pred = model(X)  # 网络输出loss = loss_fn(pred, y)  # 计算网络输出和真实值之间的差距，targets为真实值，计算二者差值即为损失# 反向传播optimizer.zero_grad()  # grad属性归零loss.backward()  # 反向传播optimizer.step()  # 每一步自动更新# 记录acc与losstrain_acc += (pred.argmax(1) == y).type(torch.float).sum().item()train_loss += loss.item()train_acc /= sizetrain_loss /= num_batchesreturn train_acc, train_lossdef test(dataloader, model, loss_fn):size = len(dataloader.dataset)  # 测试集的大小num_batches = len(dataloader)  # 批次数目test_loss, test_acc = 0, 0# 当不进行训练时，停止梯度更新，节省计算内存消耗with torch.no_grad():for imgs, target in dataloader:imgs, target = imgs.to(device), target.to(device)# 计算losstarget_pred = model(imgs)loss = loss_fn(target_pred, target)test_loss += loss.item()test_acc += (target_pred.argmax(1) == target).type(torch.float).sum().item()test_acc /= sizetest_loss /= num_batchesreturn test_acc, test_loss

# 正式训练
import copyoptimizer = torch.optim.Adam(model.parameters(), lr=1e-4)
loss_fn = nn.CrossEntropyLoss()  # 创建损失函数epochs = 50train_loss = []
train_acc = []
test_loss = []
test_acc = []best_acc = 0  # 设置一个最佳准确率，作为最佳模型的判别指标for epoch in range(epochs):# 更新学习率（使用自定义学习率时使用）# adjust_learning_rate(optimizer, epoch, learn_rate)model.train()epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)# scheduler.step() # 更新学习率（调用官方动态学习率接口时使用）model.eval()epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)# 保存最佳模型到 best_modelif epoch_test_acc > best_acc:best_acc = epoch_test_accbest_model = copy.deepcopy(model)train_acc.append(epoch_train_acc)train_loss.append(epoch_train_loss)test_acc.append(epoch_test_acc)test_loss.append(epoch_test_loss)# 获取当前的学习率lr = optimizer.state_dict()['param_groups'][0]['lr']template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}, Lr:{:.2E}')print(template.format(epoch + 1, epoch_train_acc * 100, epoch_train_loss,epoch_test_acc * 100, epoch_test_loss, lr))# 保存最佳模型到文件中
PATH = r'C:/Users/11054/Desktop/kLearning/J9_learning/best_model.pth'  # 保存的参数文件名
torch.save(model.state_dict(), PATH)print('Done')

Epoch: 1, Train_acc:73.0%, Train_loss:0.729, Test_acc:89.3%, Test_loss:0.296, Lr:1.00E-04
Epoch: 2, Train_acc:80.9%, Train_loss:0.540, Test_acc:89.8%, Test_loss:0.336, Lr:1.00E-04
Epoch: 3, Train_acc:82.0%, Train_loss:0.506, Test_acc:85.3%, Test_loss:0.486, Lr:1.00E-04
Epoch: 4, Train_acc:83.8%, Train_loss:0.443, Test_acc:87.1%, Test_loss:0.352, Lr:1.00E-04
Epoch: 5, Train_acc:85.0%, Train_loss:0.458, Test_acc:87.1%, Test_loss:0.317, Lr:1.00E-04
Epoch: 6, Train_acc:85.8%, Train_loss:0.421, Test_acc:92.0%, Test_loss:0.303, Lr:1.00E-04
Epoch: 7, Train_acc:86.7%, Train_loss:0.393, Test_acc:88.4%, Test_loss:0.335, Lr:1.00E-04
Epoch: 8, Train_acc:89.1%, Train_loss:0.314, Test_acc:94.7%, Test_loss:0.186, Lr:1.00E-04
Epoch: 9, Train_acc:88.3%, Train_loss:0.344, Test_acc:90.7%, Test_loss:0.233, Lr:1.00E-04
Epoch:10, Train_acc:88.4%, Train_loss:0.337, Test_acc:80.9%, Test_loss:0.435, Lr:1.00E-04
Epoch:11, Train_acc:89.3%, Train_loss:0.342, Test_acc:94.7%, Test_loss:0.200, Lr:1.00E-04
Epoch:12, Train_acc:90.1%, Train_loss:0.298, Test_acc:92.9%, Test_loss:0.221, Lr:1.00E-04
Epoch:13, Train_acc:89.0%, Train_loss:0.344, Test_acc:87.6%, Test_loss:0.325, Lr:1.00E-04
Epoch:14, Train_acc:90.1%, Train_loss:0.288, Test_acc:93.3%, Test_loss:0.219, Lr:1.00E-04
Epoch:15, Train_acc:91.0%, Train_loss:0.260, Test_acc:91.1%, Test_loss:0.238, Lr:1.00E-04
Epoch:16, Train_acc:89.9%, Train_loss:0.296, Test_acc:96.9%, Test_loss:0.128, Lr:1.00E-04
Epoch:17, Train_acc:90.6%, Train_loss:0.310, Test_acc:95.1%, Test_loss:0.167, Lr:1.00E-04
Epoch:18, Train_acc:89.7%, Train_loss:0.288, Test_acc:92.9%, Test_loss:0.207, Lr:1.00E-04
Epoch:19, Train_acc:92.7%, Train_loss:0.225, Test_acc:95.6%, Test_loss:0.153, Lr:1.00E-04
Epoch:20, Train_acc:92.8%, Train_loss:0.210, Test_acc:92.0%, Test_loss:0.340, Lr:1.00E-04
Epoch:21, Train_acc:92.9%, Train_loss:0.210, Test_acc:94.2%, Test_loss:0.163, Lr:1.00E-04
Epoch:22, Train_acc:92.1%, Train_loss:0.258, Test_acc:92.9%, Test_loss:0.215, Lr:1.00E-04
Epoch:23, Train_acc:91.3%, Train_loss:0.270, Test_acc:92.9%, Test_loss:0.210, Lr:1.00E-04
Epoch:24, Train_acc:92.2%, Train_loss:0.251, Test_acc:92.0%, Test_loss:0.289, Lr:1.00E-04
Epoch:25, Train_acc:93.6%, Train_loss:0.194, Test_acc:94.7%, Test_loss:0.135, Lr:1.00E-04
Epoch:26, Train_acc:95.7%, Train_loss:0.156, Test_acc:93.3%, Test_loss:0.248, Lr:1.00E-04
Epoch:27, Train_acc:94.6%, Train_loss:0.178, Test_acc:87.1%, Test_loss:0.395, Lr:1.00E-04
Epoch:28, Train_acc:93.2%, Train_loss:0.224, Test_acc:95.6%, Test_loss:0.136, Lr:1.00E-04
Epoch:29, Train_acc:94.4%, Train_loss:0.161, Test_acc:94.7%, Test_loss:0.177, Lr:1.00E-04
Epoch:30, Train_acc:94.6%, Train_loss:0.157, Test_acc:94.7%, Test_loss:0.162, Lr:1.00E-04
Epoch:31, Train_acc:92.2%, Train_loss:0.219, Test_acc:94.7%, Test_loss:0.261, Lr:1.00E-04
Epoch:32, Train_acc:94.2%, Train_loss:0.199, Test_acc:95.6%, Test_loss:0.154, Lr:1.00E-04
Epoch:33, Train_acc:94.1%, Train_loss:0.194, Test_acc:96.9%, Test_loss:0.104, Lr:1.00E-04
Epoch:34, Train_acc:95.3%, Train_loss:0.132, Test_acc:92.9%, Test_loss:0.201, Lr:1.00E-04
Epoch:35, Train_acc:95.9%, Train_loss:0.119, Test_acc:94.2%, Test_loss:0.148, Lr:1.00E-04
Epoch:36, Train_acc:94.6%, Train_loss:0.144, Test_acc:95.6%, Test_loss:0.129, Lr:1.00E-04
Epoch:37, Train_acc:97.0%, Train_loss:0.107, Test_acc:93.3%, Test_loss:0.194, Lr:1.00E-04
Epoch:38, Train_acc:96.9%, Train_loss:0.109, Test_acc:95.6%, Test_loss:0.135, Lr:1.00E-04
Epoch:39, Train_acc:95.1%, Train_loss:0.155, Test_acc:93.8%, Test_loss:0.212, Lr:1.00E-04
Epoch:40, Train_acc:94.3%, Train_loss:0.155, Test_acc:96.9%, Test_loss:0.087, Lr:1.00E-04
Epoch:41, Train_acc:94.7%, Train_loss:0.164, Test_acc:96.0%, Test_loss:0.129, Lr:1.00E-04
Epoch:42, Train_acc:94.2%, Train_loss:0.165, Test_acc:96.9%, Test_loss:0.097, Lr:1.00E-04
Epoch:43, Train_acc:96.6%, Train_loss:0.115, Test_acc:95.6%, Test_loss:0.171, Lr:1.00E-04
Epoch:44, Train_acc:96.8%, Train_loss:0.108, Test_acc:95.6%, Test_loss:0.142, Lr:1.00E-04
Epoch:45, Train_acc:95.0%, Train_loss:0.151, Test_acc:97.3%, Test_loss:0.116, Lr:1.00E-04
Epoch:46, Train_acc:94.6%, Train_loss:0.152, Test_acc:95.1%, Test_loss:0.134, Lr:1.00E-04
Epoch:47, Train_acc:96.7%, Train_loss:0.105, Test_acc:94.7%, Test_loss:0.175, Lr:1.00E-04
Epoch:48, Train_acc:97.1%, Train_loss:0.094, Test_acc:94.7%, Test_loss:0.168, Lr:1.00E-04
Epoch:49, Train_acc:96.6%, Train_loss:0.108, Test_acc:95.6%, Test_loss:0.161, Lr:1.00E-04
Epoch:50, Train_acc:96.4%, Train_loss:0.097, Test_acc:96.0%, Test_loss:0.146, Lr:1.00E-04
Done

# 结果可视化
import matplotlib.pyplot as plt
# 隐藏警告
import warningswarnings.filterwarnings("ignore")  # 忽略警告信息
plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号
plt.rcParams['figure.dpi'] = 100  # 分辨率epochs_range = range(epochs)plt.figure(figsize=(12, 3))
plt.subplot(1, 2, 1)plt.plot(epochs_range, train_acc, label='Training Accuracy')
plt.plot(epochs_range, test_acc, label='Test Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Training Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()from PIL import Imageclasses = list(total_data.class_to_idx)
print(classes)
print(total_data.class_to_idx)def predict_one_image(image_path, model, transform, classes):test_img = Image.open(image_path).convert('RGB')plt.imshow(test_img)  # 展示预测的图片test_img = transform(test_img)img = test_img.to(device).unsqueeze(0)model.eval()output = model(img)_, pred = torch.max(output, 1)pred_class = classes[pred]print(f'预测结果是：{pred_class}')# 预测训练集中的某张照片
predict_one_image(image_path=r'C:\Users\11054\Desktop\kLearning\p1_data\rain\rain1.jpg',model=model,transform=train_transforms,classes=classes)

['cloudy', 'rain', 'shine', 'sunrise']
{'cloudy': 0, 'rain': 1, 'shine': 2, 'sunrise': 3}
预测结果是：rain![在这里插入图片描述](https://i-blog.csdnimg.cn/direct/11694d2134d1499681fda128ac09c7e5.png#pic_center)

个人总结

• Inception v3 的核心思想是使用Inception模块，在同一层中并行使用不同大小的卷积核（如 1x1、3x3、5x5），以捕捉不同尺度的特征。
  通过模块化设计，网络可以在不同层次上提取多尺度的特征，从而提高模型的泛化能力
• Inception v3 将输入尺寸增大到 299 * 299，优点如下：
  捕捉更多细节：较大的输入尺寸可以保留更多的图像细节，尤其是在高分辨率图像上。
  更深的特征提取：较大的输入尺寸允许卷积层在更大的感受野内提取特征，从而提高模型的表达能力。
  推理时间的权衡：虽然增大了输入尺寸会稍微增加计算量，但 Inception v3 通过优化网络结构（如使用更小的卷积核和分解卷积操作），整体推理时间仍然可以控制在合理范围内。
• InceptionA
  InceptionA模块是Inception V3网络中第一个主要的Inception模块，它主要用于在网络的早期阶段提取大量的低级特征。InceptionA模块包含几个并行的卷积路径，这些路径的输出在深度维度上进行拼接。具体结构如下：

1x1卷积：用于降低通道数，减少计算量。
5x5卷积：通过前一个1x1卷积降维后，进行5x5卷积，提取更复杂的特征。
3x3卷积（两次）：通过1x1卷积降维后，进行两次3x3卷积，等效于一次5x5卷积，但计算量更小。
3x3最大池化：进行最大池化操作，然后通过1x1卷积恢复通道数。

• InceptionB
  InceptionB模块出现在网络的中段，用于提取中级特征。它进一步增加了网络的深度和复杂性，同时保持了计算效率。具体结构如下：

1x1卷积：用于降维和减少计算量。
1x7卷积：通过前一个1x1卷积降维后，进行1x7卷积，捕捉水平方向的特征。
7x1卷积：通过前一个1x1卷积降维后，进行7x1卷积，捕捉垂直方向的特征。
3x3最大池化：进行最大池化操作，然后通过1x1卷积恢复通道数。

• InceptionC
  InceptionC模块出现在网络的后段，用于提取高级特征。它更加复杂，使用了更多非对称的卷积核组合。具体结构如下：

1x1卷积：用于降维和减少计算量。
1x3和3x1卷积：通过前一个1x1卷积降维后，进行1x3和3x1卷积，分别捕捉不同方向的特征。
3x3最大池化：进行最大池化操作，然后通过1x1卷积恢复通道数。

• ReductionA
  ReductionA模块用于在网络的中段减少特征图的尺寸，同时增加通道数。它通过不同的路径组合来实现这一点，具体结构如下：

3x3最大池化：直接进行最大池化操作，减少特征图尺寸。
3x3卷积：通过1x1卷积降维后，进行3x3卷积，减少特征图尺寸。
两个3x3卷积（两次3x3卷积）：通过1x1卷积降维后，进行两次3x3卷积，等效于一次5x5卷积，减少特征图尺寸。

• ReductionB
  ReductionB模块用于在网络的后段进一步减少特征图的尺寸，同时增加通道数。它的结构更加复杂，具体如下：

3x3最大池化：直接进行最大池化操作，减少特征图尺寸。
3x3卷积：通过1x1卷积降维后，进行3x3卷积，减少特征图尺寸。
两个1x7和7x1卷积（两次1x7和7x1卷积）：通过1x1卷积降维后，进行两次1x7和7x1卷积，捕捉不同方向的特征，然后通过3x3卷积减少特征图尺寸。

• 辅助分支
  辅助分支是Inception V3网络中的一个特殊设计，用于在网络的中间层增加额外的监督信号，从而帮助网络更快地收敛和提高模型的鲁棒性。辅助分支的具体结构如下：

平均池化：在中间层的特征图上进行平均池化操作，减少特征图的尺寸。
1x1卷积：降维，减少通道数。
全连接层：将降维后的特征图展平，通过全连接层生成辅助分类器的输出。
辅助分类器：用于计算中间层的损失，这个损失在训练过程中会被加到最终的损失函数中。