Piotr Żelasko
Piotr Żelasko
> Currently we have to patch Nemo's Lhotse to add it to sampler which is very hard to maintain. Could you make a PR with this feature?
If you set drop_last is to False it will return a single batch with all available data in your dataset per epoch.
Yes, you'd need to set it separately in every subprocess. For PT dataloader there's a worker_init_fn you can provide.
If you don't have the time to implement the adjustment, we can merge this as-is instead.
You might just get lucky if you set `lhotse.set_dill_enabled(True)` somewhere in your code (or `LHOTSE_DILL_ENABLED=1`). Otherwise you'll have to try and debug where this generator object is created. I don't...
Nobody contributed this recipe to lhotse AFAIK
Use `lhotse.qa.fix_manifests` on the imported supervisions + recordings.
Hi @kerolos, thanks for opening this discussion! I can help you get your video recipe set up. > Is there a plan to add a recipe that supports video data...
> * How can avoid this error from the class Recording ? Video loading features depend on you having a recent version of pytorch, torchaudio, and compatible ffmpeg version to...
As for your other question: > Option 1: Saving Cuts from Recordings and Supervisions Manifests (from_jsonl): Yeah we'll need to add mp4 support for AudioTarWriter. I don't have the bandwidth...