視訊位址:[https://www.bilibili.com/video/BV1Y7411d7Ys?p=5]
實作步驟:
1.準備資料集
2.設計資料模型,前向傳播計算預測值
3.構造損失函數(目的為反向傳播)和優化器(更新梯度)
4.訓練周期(前饋、回報、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__()
# 第一個參數輸入次元 第二個參數輸出次元
self.linear = torch.nn.Linear(1, 1)
# 前饋
def forward(self, x):
# 繼承 _call_(),實際上調用
y_pred = self.linear(x)
return y_pred
model = LinearModel()
criterion = torch.nn.MSELoss(reduction='sum')
# 優化器 model.parameters()擷取模型中需要優化的參數
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
# 訓練
for epoch in range(100):
# 前饋
y_pred = model(x_data)
# 損失
loss = criterion(y_pred, y_data)
print(epoch, loss.item())
# 梯度清零
optimizer.zero_grad()
# 反向傳播
loss.backward()
# 更新優化參數
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)