Nay San

+1 to issue. All I really need is `tqdm` to work (see https://github.com/jupyter/docker-stacks/issues/1715) and am torn because that this alone isn't enough of a reason to switch to notebooks but...

Hi @pzelasko — I was wondering whether community contributions would still be welcome for some of the following (listed above), e.g.: - Using Lhotse with PytorchLightning - Determining the optimal...

Hi @mrezasoltani — I think since it's not a straightforward thing to get the length of a BucketingSampler, I suggest using a global progress bar based on `MAX_STEPS`. ```python from...

1. TokenCollator > As a minor technical note, I don't think you need any of // (and actually even ) for CTC models. Ah, I see — thanks @pzelasko !...

Hi @mrezasoltani — Here's some dummy code of my 'hack' around, taking advantage of two facts: - For training sampler, since it will be shuffled at each epoch and it...

**Progress bar** I searched `IterableDataset` inside the PyTorch Lightning repo. It looks like if you switch to using a `RichProgressBar` (instead of the default `tqdm`) they have some 'support' for...