Pytorch 中的F范数写法

f范数

criterion = torch.nn.MSELoss(reduction='mean')

1,2范数

coef_q = torch.norm(Q,p =1,dim =1)

loss_q = torch.norm(coef_q,p =2)