Closed splinter21 closed 2 years ago
https://github.com/chengdazhi/Deformable-Convolution-V2-PyTorch/issues/9#issuecomment-457592960
According to this, 1/10 lr->stabler for training in ImageNet VID, but no need for training in coco
Is it needed for VSR?
i think so,dcn is not stable
https://github.com/chengdazhi/Deformable-Convolution-V2-PyTorch/issues/9#issuecomment-457592960
According to this, 1/10 lr->stabler for training in ImageNet VID, but no need for training in coco
Is it needed for VSR?