07-pytorch(模型的保存与提取)

import torch
from torch.autograd import Variable

# fake data
x = torch.unsqueeze(torch.linspace(-1,1,100),dim=1)
y = x.pow(2) + 0.2*torch.rand(x.size())
x,y = Variable(x),Variable(y)

保存数据

全部保留
torch.save(net1 ,'net.pkl')
参数保留
torch.save(net1.state_dict(), 'net_params.pkl' )
提取神经网络
net2 = torch.load('.//pkl//net.pkl')
用参数还原神经网络
net3 = torch.nn.Sequential(
torch.nn.Linear(1,10),
torch.nn.ReLu(),
torch.nn.Linear(10,1)
)
net3.load_state_dict(torch.load('.//pkl/net_params'))

def save():
    # save net1
    net1 = torch.nn.Sequential(
            torch.nn.Linear(1,10),
            torch.nn.ReLU(),
            torch.nn.Linear(10,1)
    )
    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()
    ### 保存模型
    torch.save(net1 ,'net.pkl')
    ### 只保留参数
    torch.save(net1.state_dict(),  'net_params.pkl' )

def restore_net():
    # 提取全部的神经网络
    net2 = torch.load('.//pkl//net.pkl')
def restore_params():
    # 使用参数复原神经网络
    net3 = torch.nn.Sequential(
            torch.nn.Linear(1,10),
            torch.nn.ReLu(),
            torch.nn.Linear(10,1)
    )
    net3.load_state_dict(torch.load('.//pkl/net_params'))

# save()
restore_net()