Everything works with 1 GPU and num_workers > 0, but if the number of GPUs is set to > 1 I get this error:
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0! (when checking argument for argument mat1 in method wrapper_addmm)
Everything works with 1 GPU and
num_workers
> 0, but if the number of GPUs is set to > 1 I get this error:See these guides as starting point: