Closed shawnl111 closed 5 months ago
have the same problem. if num_worker>0 and batch_size>1, I have the same error. if num_worker = 0 and batch_size>1, stack error.
num_worker = 0 and batch_size=1, it is ok
The dataset itself has tensors of different shapes. So they can't be stacked anyway, in my opinion.
training on torch2.2.1 and cuda11.8 on widows11