Pytorch 感知机 痛定思痛。 2021-11-17 06:20 321阅读 0赞 示例代码： import torch from torch.nn import functional as F x = torch.randn(1, 10) w = torch.randn(1, 10, requires_grad=True) # 对输出用sigmoid激活 o = torch.sigmoid(x @ w.t()) print("输出值:", o) # 计算MSE loss = F.mse_loss(torch.ones(1, 1), o) print("损失:", loss) # 计算梯度 loss.backward() print("损失对w的导数:", w.grad) 输出结果： 输出值: tensor([[0.9352]], grad_fn=<SigmoidBackward>) 损失: tensor(0.0042, grad_fn=<MeanBackward1>) 损失对w的导数: tensor([[-0.0051, -0.0106, 0.0006, -0.0090, -0.0068, -0.0035, 0.0017, 0.0010, -0.0099, 0.0064]])
还没有评论，来说两句吧...