LeNet实践

import torch
from torch import nn
from d2l import torch as d2l

net = nn.Sequential(
    nn.Conv2d(1, 6, kernel_size=5, padding=2), nn.LeakyReLU(0.1),
    nn.MaxPool2d(kernel_size=2, stride=2),
    nn.Conv2d(6, 16, kernel_size=5), nn.LeakyReLU(0.1),
    nn.MaxPool2d(kernel_size=2, stride=2),
    nn.Flatten(),
    nn.Linear(16 * 5 * 5, 120), nn.LeakyReLU(0.1),
    nn.Linear(120, 84), nn.ReLU(),
    nn.Linear(84, 10))

X = torch.rand(size=(1, 1, 28, 28), dtype=torch.float32)
for layer in net:
    X = layer(X)
    print(layer.__class__.__name__,'output shape: \t',X.shape)

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size=batch_size) 

def evaluate_accuracy_gpu(n