About CondInst
wudaxianzi opened this issue · 0 comments
wudaxianzi commented
When evaluating，comes error non-empty 3D or 4D(batch mode)tensor expected for input, but got: [ torch.cuda.FloatTenson{09,1,209,288} ，I know it's about some images doesn't have mask.So I wanna know where could I modify to skip these images to make evaluating goes well.Thanks a lot if you can help me~