Matthew Leigh

Sinkhorn loss always renables gradient tracking

1

Essentially whenever I calculate the sinkhorn loss, pytorch's internal gradient tracking turns back on. I believe this is to do with line 280 in geomloss/sinkhorn_divergence.py: `torch.autograd.set_grad_enabled(True)` This is obviously not...

IterableDataset with CORRECT length causes validation loop to be skipped

9

### Bug description This is related to this issue: https://github.com/Lightning-AI/pytorch-lightning/issues/10290 Whereby an IterableDataset with a length defined wont trigger a validation epoch, even if the defined length is correct so...