Warning: Grad strides do not match bucket view strides.

[botaoye]

Hi, thanks for your wonderful work. When I use Xcit for another task as the backbone, it comes the warning of Warning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed. It is caused by feeding incontiguous tensors to view-style operators. While I can find some positions that cause this warning, there seem exits several different code lines which can cause this warning and I failed to find all of them. I wonder if you have also encountered this warning and do you have any advice to solve this problem?