white_hair_的博客

import torchimport torch.nn as nnimport torch.nn.functional as Fimport torchvisionimport torchvision.transforms as transformsimport torch.utils.dataimport torch.optim as optimdevice = torch.device("cuda:0" if torch.cuda.is_available() else "cpu").

主要实现的部分一、借助nn.Module建立一个神经网络二、建立此网络的可学习的参数列表三、引用MSEloss作为损失函数四、梯度的反向传播和更新 更新规则的设定import torchimport torch.nn as nnimport torch.nn.functional as Fimport torch.optim as optim # SGD, Nesterov-SGD, Adam,RMSProp的权重更新规则class Net(nn.Module): #.