Pytorch常用的API

Pytorch常用API

1. nn.Module

   a. _init_

   b. forward 完成一次向前计算的过程

nn.Linear(input的特征数量，输出的特征数量)

nn.Module定义了__call__方法，实现的就是调用forward方法

实例化模型，传入数据，计算结果

2. 优化器类

   a. torch.optim.SGD(参数，学习率)

   b. torch.optim.Adam(参数，学习率)

参数的获取model.parameters()，获取模型中所有requires_grad=True的参数

3. Loss函数

   loss_fn=nn.MSEloss()

import torch.nn as nn
import torch.optim as optim

device=torch.device('cuda' if torch.cuda.is_available() else 'cpu')
class myLinear(nn.Module):
    def __init__(self):
        super(myLinear,self).__init__()
        self.linear=nn.Linear(1,1)
    def forward(self, x):
        return self.linear(x)
net=myLinear().to(device)
loss_fn=nn.MSELoss()
optimizer=optim.SGD(net.parameters(),0.001)
print(net)
print(loss_fn)
print(optimizer)

myLinear(
  (linear): Linear(in_features=1, out_features=1, bias=True)
)
MSELoss()
SGD (
Parameter Group 0
    dampening: 0
    lr: 0.001
    momentum: 0
    nesterov: False
    weight_decay: 0
)

x=torch.randn(500,1).to(device)
y_true=2*x+1

for iter in range(5000):
    out=net(x)
    loss=loss_fn(out,y_true)
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()
    if iter%100==0:
        print(loss.item(),list(net.parameters()))

1.8090484665833628e-09 [Parameter containing:
tensor([[ 2.0000]], device='cuda:0'), Parameter containing:
tensor([ 1.0000], device='cuda:0')]