XQ

In the pytorch-lightning2.2.1 environment, Maml is implemented by wrapping the meta_learn function with torch.enable_grad, which makes the model's outer losses uncomputed during validation, resulting in memory overflow. This can be...

MiniImagenet in your points to https://github.com/hytseng0509/CrossDomainFewShot file download links seems to fail, to open the show 404, excuse me whether can provide the data

Excellent work, and supporting Flash Attention will make it more widely used.

Because of the encapsulation of cal_inf_loss, it is not very convenient to obtain prediction labels, it is recommended that the author add the function encapsulation of obtaining prediction labels and...