基本上所有的网络训练过程都可以按这种过程来实现
这里以简单的LeNet-5识别手写体数字为例,让大家看到训练的完整流程:
import torch
from torchvision import datasets
from torchvision import transforms
from torch.utils.data import DataLoader
import torch.nn as nn
import torch.optim as optim
# 设置超参数
batch_size = 64
epochs = 40
lr = 0.001
# 1、数据
train_dataset = datasets.MNIST(root='./data',train=True,transform=transforms.ToTensor(),download=True)
test_dataset = datasets.MNIST(root='./data',train=False,transform=transforms.ToTensor())
train_loader = DataLoader(train_dataset,batch_size=batch_size,shuffle=True)
train_loader = DataLoader(test_dataset,batch_size=batch_size,shuffle=False)
# 2、模型
class LeNet5(nn.Module):
def __init__(self, in_dim, n_class):
super(LeNet5, self).__init__()
self.conv = nn.Sequential(
nn.Conv2d(in_dim, 6, 5, stride=1, padding=2),
nn.ReLU(True),
nn.MaxPool2d(2, 2),
nn.Conv2d(6, 16, 5, stride=1, padding=0),
nn.ReLU(True),
nn.MaxPool2d(2, 2))
self.fc = nn.Sequential(
nn.Linear(400, 120),
nn.Linear(120, 84),
nn.Linear(84, n_class))
def forward(self, x):
out = self.conv(x)
out = out.view(out.size(0), -1)
out = self.fc(out)
return out
# 实例化模型
model = LeNet5(1,10)
# 3、损失函数
criterion = nn.CrossEntropyLoss()
# 4、优化器
optimizer = optim.SGD(model.parameters(),lr=lr)
# 5、迭代训练
def train(model,optimizer,criterion,train_loader,num_epochs):
model.train() # 声明这里的model是作为训练
for epoch in range(num_epochs):
train_loss = 0.0
train_acc= 0.0
train_len = 0.0
total = 0.0
for i,data in enumerate(train_loader):
# 前向传播
img,label = data
output = model(img)
# 反向传播
optimizer.zero_grad()
loss = criterion(output,label)
loss.backward()
# 更新参数
optimizer.step()
train_loss += loss.item()
train_len += output.shape[0]
# 准确率
pred = torch.argmax(output,1)
num_correct = pred.eq(label.data.view_as(pred)).sum()
train_acc += num_correct.item()
total += label.size(0)
print('Finish {} epoch, Loss: {:.6f}, Acc: {:.6f}'.format(
epoch + 1, train_loss / (train_len), train_acc / (total)))
if __name__ == '__main__':
train(model,optimizer,criterion,train_loader,num_epochs=epochs)
# 6、保存模型
torch.save(model.state_dict(), './lenet5.pth')
后续将会实现其他经典网络实战案例,如果对你有帮助的话可以点赞收藏关注