实验代码
import torch
import torch.nn as nn
#y = wx + b
class MyModel(nn.Module):
def __init__(self):
super(MyModel,self).__init__()
#自定义代码
# self.w = torch.rand([500,1],requires_grad=True)
# self.b = torch.tensor(0,dtype=torch.float,requires_grad=True)
# self.lr = nn.Linear(1,1)
self.lr1 = nn.Linear(1,10)
# self.lr2 = nn.Linear(10,20)
# self.lr3 = nn.Linear(20,1)
def forward(self,x): #完成一次前项计算
# y_predict = x*self.w + self.b
# return y_predict
# return self.lr(x)
out1 = self.lr1(x)
# out2 = self.lr2(out1)
# out = self.lr3(out2)
return out1
if __name__ == '__main__':
model = MyModel()
# print(model.parameters())
for i in model.parameters():
print(i)
print("*"*100)
# y_predict = model(torch.FloatTensor([10]))
# print(y_predict)
Linear实现线性回归,cuda版本
import torch
import torch.nn as nn
from torch import optim
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
class MyModel(nn.Module):
def __init__(self):
super(MyModel,self).__init__()
self.lr = nn.Linear(1,1)
def forward(self,x):
return self.lr(x)
#准备数据 如果使用cuda,数据和模型需要to(device)
x = torch.rand([500,1]).to(device)
y_true = 3*x + 0.8
#实例化模型
model = MyModel().to(device)
#实例化优化器
optimizer = optim.Adam(model.parameters(),lr=0.1)
#实例化损失函数
loss_fn = nn.MSELoss()
for i in range(500):
#梯度置零
optimizer.zero_grad()
#调用模型得到预测值
y_predict = model(x)
#损失函数,计算损失
loss = loss_fn(y_predict,y_true)
#反向传播计算梯度
loss.backward()
#更新参数
optimizer.step()
#打印部分数据
if i%10 ==0:
print(i,loss.item())
for param in model.parameters():
print(param.item())