Is there any support Speaker diarisation

[sh1man]

Will support be added for diarization?

[csukuangfj]

We have already supported it in sherpa-onnx

Please see
https://k2-fsa.github.io/sherpa/onnx/speaker-diarization/index.html

[sh1man]

onnx diarization can be deployed in triton ?
example
https://github.com/k2-fsa/sherpa/tree/master/triton/whisper

We have already supported it in sherpa-onnx

Please see https://k2-fsa.github.io/sherpa/onnx/speaker-diarization/index.html

[csukuangfj]

no,.not yet

[sh1man]

Is there a complete example with ASR + Diarization integration to prod

[csukuangfj]

no

[yuekaizhang]

onnx diarization can be deployed in triton ? example https://github.com/k2-fsa/sherpa/tree/master/triton/whisper

If native pytorch fp16 model could fulfill your needs, you may try the triton python backend to deploy it. See https://github.com/triton-inference-server/tutorials/blob/main/HuggingFace/python_model_repository/python_vit/1/model.py.

For onnx models, you may need to make some modifications.