speaker diarization custom dataset creation

[arrrrr3186]

Hello @hbredin,

I'm currently encountering issues with the performance of a speaker diarization model and I'm considering creating my own dataset to improve its results. Could anyone provide guidance on the steps and tools required to create a custom dataset for this purpose?

Thank you!

[arrrrr3186]

Hello @hbredin,

if you can guide me on this, that would be a great help to me.

Regards,

[FrenchKrab]

I assume you have an existing set of audio data and associated annotations that you want to finetune your model on.
Please take a look at the resources in this comment to set it up with pyannote:
#180 (comment)