Piotr Żelasko

I think that we should consider a refactoring of snowfall to share more code between the recipes. The number of similar issues will grow exponentially as we start adding new...

There is also a transform called [ExtraPadding](https://github.com/lhotse-speech/lhotse/blob/master/lhotse/dataset/cut_transforms/extra_padding.py) that adds a fixed number N of padding frames to the cut (N/2 on each side); I can extend it so that it...

FYI if we train full Libri with the current augmentation setup for 10 epochs on 4 x 32GB V100 with 4 x learning rate, average last 3 epochs and use...

I’m happy to do it, it’s actually been on my radar for some time now, but my time is also spread a bit thin lately. Anyway, I guess let’s just...

Actually I just realized it will be interesting for you to see more info; here's a WER breakdown by epoch and numbers for different average settings, and with/without rescoring. You...

Also, I think 1 epoch takes only 2-2.5 hours in this setting (I don't know exactly because I still haven't figured out how to fix the logs other than with...

I think I did it concurrently to your question -- only the last result in my previous message has the rescoring turned on.

> > Also, I think 1 epoch takes only 2-2.5 hours in this setting (I don't know exactly because I still haven't figured out how to fix the logs other...

@glynpu sure! You should be able to download it here: https://livejohnshopkins-my.sharepoint.com/:f:/g/personal/pzelask2_jh_edu/EjpFSUZ1WXlItIWlf-YemmIBTbNkbA3fovl_kZv0tQFupw?e=JZHh6x LMK if that doesn't work.

I think “best_model” doesn’t store the optimizer, scheduler, etc. state dicts for resuming training. Also it is not necessarily best, since it’s picked based on dev loss and not on...