LeNet
LeNet
一.网络架构
利用Sequential块实现Lenet
import torch
from torch import nn
#利用Sequential实现Lenet
net=nn.Sequential(
nn.Conv2d(1,6,5,padding=2),
nn.Sigmoid(),
nn.AvgPool2d(kernel_size=2,stride=2),
nn.Conv2d(6,16,kernel_size=5),
nn.Sigmoid(),
nn.AvgPool2d(kernel_size=2,stride=2),
nn.Flatten(),
nn.Linear(16*5*5,120),
nn.Sigmoid(),
nn.Linear(120,84),
nn.Sigmoid(),
nn.Linear(84,10)
)
X=torch.rand((1,1,28,28))#表示 1 个样本、1 个通道、28x28 像素的图像
for layer in net:
X=layer(X)
print(layer.__class__.__name__,"output_shape\t",X.shape)获得每一层名称和大小
二.利用Lenet网络实现数字0~9分类
完整训练代码
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.Sigmoid(),
nn.AvgPool2d(kernel_size=2, stride=2),
nn.Conv2d(6, 16, kernel_size=5), nn.Sigmoid(),
nn.AvgPool2d(kernel_size=2, stride=2),
nn.Flatten(),
nn.Linear(16 * 5 * 5, 120), nn.Sigmoid(),
nn.Linear(120, 84), nn.Sigmoid(),
nn.Linear(84, 10))
batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size=batch_size)
def evaluate_accuracy_gpu(net, data_iter, device=None): #@save
"""使用GPU计算模型在数据集上的精度"""
if isinstance(net, nn.Module):
net.eval() # 设置为评估模式
if not device:
device = next(iter(net.parameters())).device
# 正确预测的数量，总预测的数量
metric = d2l.Accumulator(2)
with torch.no_grad():
for X, y in data_iter:
if isinstance(X, list):
# BERT微调所需的（之后将介绍）
X = [x.to(device) for x in X]
else:
X = X.to(device)
y = y.to(device)
metric.add(d2l.accuracy(net(X), y), y.numel())
return metric[0] / metric[1]
#@save
def train_ch6(net, train_iter, test_iter, num_epochs, lr, device):
"""用GPU训练模型(在第六章定义)"""
def init_weights(m):
if type(m) == nn.Linear or type(m) == nn.Conv2d:
nn.init.xavier_uniform_(m.weight)
net.apply(init_weights)
print('training on', device)
net.to(device)
optimizer = torch.optim.SGD(net.parameters(), lr=lr)
loss = nn.CrossEntropyLoss()
animator = d2l.Animator(xlabel='epoch', xlim=[1, num_epochs],
legend=['train loss', 'train acc', 'test acc'])
timer, num_batches = d2l.Timer(), len(train_iter)
for epoch in range(num_epochs):
# 训练损失之和，训练准确率之和，样本数
metric = d2l.Accumulator(3)
net.train()
for i, (X, y) in enumerate(train_iter):
timer.start()
optimizer.zero_grad()
X, y = X.to(device), y.to(device)
y_hat = net(X)
l = loss(y_hat, y)
l.backward()
optimizer.step()
with torch.no_grad():
metric.add(l * X.shape[0], d2l.accuracy(y_hat, y), X.shape[0])
timer.stop()
train_l = metric[0] / metric[2]
train_acc = metric[1] / metric[2]
if (i + 1) % (num_batches // 5) == 0 or i == num_batches - 1:
animator.add(epoch + (i + 1) / num_batches,
(train_l, train_acc, None))
test_acc = evaluate_accuracy_gpu(net, test_iter)
animator.add(epoch + 1, (None, None, test_acc))
print(f'loss {train_l:.3f}, train acc {train_acc:.3f}, '
f'test acc {test_acc:.3f}')
print(f'{metric[2] * num_epochs / timer.sum():.1f} examples/sec '
f'on {str(device)}')
lr, num_epochs = 0.9, 10
train_ch6(net, train_iter, test_iter, num_epochs, lr, d2l.try_gpu())输出的结果
关键代码分析
（1）Lenet网络
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.Sigmoid(),
nn.AvgPool2d(kernel_size=2, stride=2),
nn.Conv2d(6, 16, kernel_size=5), nn.Sigmoid(),
nn.AvgPool2d(kernel_size=2, stride=2),
nn.Flatten(),
nn.Linear(16 * 5 * 5, 120), nn.Sigmoid(),
nn.Linear(120, 84), nn.Sigmoid(),
nn.Linear(84, 10))
（2）传入数据集
batch_size = 256#批量大小，每次训练样本数
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size=batch_size)（3）计算准确率
def evaluate_accuracy_gpu(net, data_iter, device=None): #device指cpu或gpu
"""使用GPU计算模型在数据集上的精度"""
if isinstance(net, nn.Module):
net.eval() # 设置为评估模式
if not device:
device = next(iter(net.parameters())).device#如果没有自定义在什么设备上跑，就默认在网络的第一个参数运行的设备上跑
# 正确预测的数量，总预测的数量，所以是2（2个参数）
metric = d2l.Accumulator(2)
with torch.no_grad():
for X, y in data_iter:
if isinstance(X, list):#如果是列表型
# BERT微调所需的（之后将介绍）
X = [x.to(device) for x in X]
else:
X = X.to(device)
y = y.to(device)
metric.add(d2l.accuracy(net(X), y), y.numel())#正确的样本数量d2l.accuracy(net(X), y) ，总样本数y.numel()
return metric[0] / metric[1]# metric = d2l.Accumulator(2)，第一个参数metric[0] 正确样本数，第二个参数 metric[1]总样本数
（4）核心函数
def train_ch6(net, train_iter, test_iter, num_epochs, lr, device):
"""用GPU训练模型(在第六章定义)"""
def init_weights(m):
if type(m) == nn.Linear or type(m) == nn.Conv2d:#全连接层或卷积层
nn.init.xavier_uniform_(m.weight)#Xavier 均匀分布 进行初始化
net.apply(init_weights)
print('training on', device)
net.to(device)
optimizer = torch.optim.SGD(net.parameters(), lr=lr)#优化器，torch.optim.SGD：指定使用随机梯度下降（Stochastic Gradient Descent） 优化器，这是最基础的优化器之一。
loss = nn.CrossEntropyLoss()#损失函数，常用于分类
animator = d2l.Animator(xlabel='epoch', xlim=[1, num_epochs],
legend=['train loss', 'train acc', 'test acc'])#绘图
timer, num_batches = d2l.Timer(), len(train_iter)
for epoch in range(num_epochs):
# 训练损失之和，训练准确率之和，样本数
metric = d2l.Accumulator(3)
net.train()#训练
for i, (X, y) in enumerate(train_iter):
timer.start()
optimizer.zero_grad()
X, y = X.to(device), y.to(device)#放入gpu
y_hat = net(X)#前向传播
l = loss(y_hat, y)
l.backward()#反向传播
optimizer.step() # 更新参数
with torch.no_grad():
#l * X.shape[0]损失, d2l.accuracy(y_hat, y)正确样本数, X.shape[0]总样本数
metric.add(l * X.shape[0], d2l.accuracy(y_hat, y), X.shape[0])
timer.stop()
train_l = metric[0] / metric[2]#损失率
train_acc = metric[1] / metric[2]#准确率
if (i + 1) % (num_batches // 5) == 0 or i == num_batches - 1:#绘图
animator.add(epoch + (i + 1) / num_batches,
(train_l, train_acc, None))
test_acc = evaluate_accuracy_gpu(net, test_iter)#test_iter测试数据量
animator.add(epoch + 1, (None, None, test_acc))
print(f'loss {train_l:.3f}, train acc {train_acc:.3f}, '
f'test acc {test_acc:.3f}')
print(f'{metric[2] * num_epochs / timer.sum():.1f} examples/sec '#样本速率
f'on {str(device)}')
（5）运行程序
lr, num_epochs训练几代 = 0.9, 10#lr学习率， num_epochs训练几代
train_ch6(net, train_iter, test_iter, num_epochs, lr, d2l.try_gpu())
浙公网安备 33010602011771号