Only for first batch - Was asked to gather along dimension 0, but all input tensors were scalars

Hi guys,

I know that there are similar issues, but it seems a little bit different from my case

I tried to train my model with the dataparallel with multi-GPUs
but, I got this warning only for the first batch among total iterations.

Does anyone know how to suppress this warning? or is it okay to ignore this?

thx