1.保存神经网络

　　速度较慢

2.只保存神经网络参数

　　速度快，这种方式将会提取所有的参数, 然后再放到你的新建网络中

代码：

import torch
import matplotlib.pyplot as plt
import torch.nn.functional as F     # 激励函数都在这

torch.manual_seed(1)    # reproducible

# 假数据 用了torch.manual_seed(1)所以假数据是固定不变的
x = torch.unsqueeze(torch.linspace(-1, 1, 100), dim=1)  # x data (tensor), shape=(100, 1)
y = x.pow(2) + 0.2*torch.rand(x.size())  # noisy y data (tensor), shape=(100, 1)
def save():
    #构建神经网络
    net1=torch.nn.Sequential(
        torch.nn.Linear(1,10),
        torch.nn.ReLU(),
        torch.nn.Linear(10,1)
    )
    #训练
    optimizer = torch.optim.SGD(net1.parameters(),lr=0.5)
    loss_func=torch.nn.MSELoss()
    for t in range(100):
        prediction=net1(x)
        loss=loss_func(prediction,y)
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
    #保存
    torch.save(net1, 'net.pkl')#整个网络
    torch.save(net1.state_dict(),'net_params.pkl')#网络的参数
def restore_net():
    # restore entire net1 to net2
    net2 = torch.load('net.pkl')
    prediction = net2(x)
def restore_params():
    # 新建 net3
    net3 = torch.nn.Sequential(
        torch.nn.Linear(1, 10),
        torch.nn.ReLU(),
        torch.nn.Linear(10, 1)
    )

    # 将保存的参数复制到 net3
    net3.load_state_dict(torch.load('net_params.pkl'))
    prediction = net3(x)
save()
restore_net()
restore_params()

输出图：