import torch
import matplotlib.pyplot as plt
learning_rate = 0.1

#准备数据 #y = 3x +0.8
x = torch.randn([500,1])
y_true = 3*x + 0.8

#计算预测值
w = torch.rand([],requires_grad=True)
b = torch.tensor(0,dtype=torch.float,requires_grad=True)

for i in range(50):
    #梯度默认会累加,梯度手动清零
    for j in [w,b]:
        if j.grad is not None:
            j.grad.data.zero_()
    y_predict = x*w +b
    #计算损失
    loss = (y_predict-y_true).pow(2).mean()
    loss.backward()
    #更新参数
    w.data = w.data - learning_rate * w.grad
    b.data = b.data - learning_rate * b.grad
    print(i,loss.item())
    print(w.data,b.data)

plt.figure(figsize=(20,8))
plt.scatter(x.numpy(),y_true.numpy())

y_predict = x*w + b
plt.plot(x.numpy(),y_predict.detach().numpy(),c="red")
plt.show()

  pytorch实现手动线性回归