batch_videos = torch.stack(batch_videos, dim=0).permute(0,1,4,2,3) #B,T,C,H,W for spatial augmentation
RuntimeError: stack expects each tensor to be equal size, but got [64, 240, 260, 3] at entry 0 and [64, 400, 720, 3] at entry 1
But input shape was different. How can I fix this?
I tried to train on WLASL dataset.
batch_videos = torch.stack(batch_videos, dim=0).permute(0,1,4,2,3) #B,T,C,H,W for spatial augmentation RuntimeError: stack expects each tensor to be equal size, but got [64, 240, 260, 3] at entry 0 and [64, 400, 720, 3] at entry 1
But input shape was different. How can I fix this?