Hervé BREDIN

icon indicating a website link herve.niderb.fr/

icon company @pyannoteai and CNRS icon location France icon location The "pyannote" guy. Co-Founder and CSO of @pyannoteai. Currently on leave from CNRS

Results 270 comments of Hervé BREDIN

save speech separation results to disk throw IndexError，size of diarization.labels() and shape of sources.data is not same

Hey @joonaskalda, I believe @RioLLee correctly pin-pointed the reason why this happens. Do you think you'll find time to have a look?

Upgrade to pyannote-audio 4

That's great @borgoat. FYI, I just released version `4.0.1` of `pyannote.audio` that fixes support for pyannoteAI premium diarization models. Therefore, once you update this PR to use `4.0.1`, running the...

Upgrade to pyannote-audio 4

> * Right now we're using the HF token for pyannote too - I guess that's incorrect @hbredin assuming one may want to use pyannote.ai? Using `--hf_token {pyannoteAI-api-key}` should work...

When's the next release? (for numpy 2.0 compatibility)

No ETA unfortunately, due to limit resources on our side. Note than you can install pyannote from the develop branch source directly.

When's the next release? (for numpy 2.0 compatibility)

3.3.2 has just been released with numpy 2.x fix.

When fine tuning pretrained segmentation model using pyannote.audio==3.1.1 on well-defined and registered custom finance dataset, it shows the following error message 'PicklingError: Can't pickle <class 'pyannote.database.registry.Finance'>: attribute lookup Finance on pyannote.database.registry failed'

Training on Windows is not supported.

When fine tuning pretrained segmentation model using pyannote.audio==3.1.1 on well-defined and registered custom finance dataset, it shows the following error message 'PicklingError: Can't pickle <class 'pyannote.database.registry.Finance'>: attribute lookup Finance on pyannote.database.registry failed'

Unfortunately, I have no easy way to check whether any of these would work in a Virtualbox. Maybe someone who did that already in pyannote user base can chime in?

Fix numpy.NAN which is numpy.nan

Thanks! 🎉 Will be part of next release.

fix: fix type hint of Timeline.to_annotation

Closing in favor of #96

pep 420: support native namespace packages

Happy to review PRs (on every pyannote.* projects) adding this feature!