论文要点：VGG（模型M+公式F+代码C）

# VGG

论文： very deep convolutional networks for large-scale image recognition by Karen Simonyan, et al

1. 模型：

图片来源：动手学深度学习

2. 代码：

# 代码来自动手学深度学习 by 李沐
import torch
from torch import nn
from d2l import torch as d2l

def vgg_block(num_convs, in_channels, out_channels):
    layers = []
    for _ in range(num_convs):
        layers.append(nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1))
        layers.append(nn.ReLU())
        in_channels = out_channels
    layers.append(nn.MaxPool2d(2, stride=2))
    return nn.Sequential(*layers)

conv_arch = ((1, 64), (1, 128), (2, 256), (2, 512), (2, 512))

def vgg_my(conv_arch):
    conv_blocks = []
    in_channels = 1
    for num_convs, out_channels in conv_arch:
        conv_blocks.append(vgg_block(num_convs, in_channels, out_channels))
        in_channels = out_channels
    return nn.Sequential(*conv_blocks, nn.Flatten(), 
                         nn.Linear(out_channels * 7 * 7, 4096), nn.ReLU(), nn.Dropout(0.5),
                         nn.Linear(4096, 4096), nn.ReLU(), nn.Dropout(0.5),
                         nn.Linear(4096, 10)
                        )

X = torch.rand((1, 1, 224, 224))  #高宽减半，通道数翻倍
for layer in vgg_my(conv_arch):
    X = layer(X)
    print(layer.__class__.__name__,":", X.shape)

net = vgg_my(conv_arch)

ratio = 4  #因通道数过大，故设计一个较小的VGG
small_conv_arch = [(pair[0], pair[1] // ratio) for pair in conv_arch]
net = vgg_my(small_conv_arch)

lr, num_epochs, batch_size = 0.05, 10, 128
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size, resize=224)
d2l.train_ch6(net, train_iter, test_iter, num_epochs, lr, d2l.try_gpu())