Diarization / unknown speaker

[renaudcerrato]

Is it possible to identify unknown speakers (i.e without registering) and get something like speaker_1, speaker_2, speaker_3?

[csukuangfj]

Please see our example.at

https://github.com/k2-fsa/sherpa-onnx/blob/master/python-api-examples/speaker-identification-with-vad-dynamic.py

[renaudcerrato]

Thanks! Will take a look!

[csukuangfj]

@renaudcerrato

We have supported speaker diarization.

[renaudcerrato]

Thanks @csukuangfj! I didn't had the time yet to try to play with. What do you mean? Did you pushed new examples recently?

[csukuangfj]

Please see the usage example at

#1396 (comment)

More APIs for different languages are being added.