视频地址:[https://www.bilibili.com/video/BV1Y7411d7Ys?p=5]
实现步骤:
1.准备数据集
2.设计数据模型,前向传播计算预测值
3.构造损失函数(目的为反向传播)和优化器(更新梯度)
4.训练周期(前馈、反馈、更新)
import torch
x_data = torch.Tensor([[1.0], [2.0], [3.0]])
y_data = torch.Tensor([[2.0], [4.0], [6.0]])
class LinearModel(torch.nn.Module):
# 初始化
def __init__(self):
super(LinearModel, self).__init__()
# 第一个参数输入维度 第二个参数输出维度
self.linear = torch.nn.Linear(1, 1)
# 前馈
def forward(self, x):
# 继承 _call_(),实际上调用
y_pred = self.linear(x)
return y_pred
model = LinearModel()
criterion = torch.nn.MSELoss(reduction='sum')
# 优化器 model.parameters()获取模型中需要优化的参数
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
# 训练
for epoch in range(100):
# 前馈
y_pred = model(x_data)
# 损失
loss = criterion(y_pred, y_data)
print(epoch, loss.item())
# 梯度清零
optimizer.zero_grad()
# 反向传播
loss.backward()
# 更新优化参数
optimizer.step()
print('w=', model.linear.weight.item())
print('b=', model.linear.bias.item())
x_test = torch.Tensor([[4.0]])
y_test = model(x_test)
print('y_pred=', y_test.data)