Request to add sherpa-ncnn and sherpa-onnx

[csukuangfj]

They both support real-time speech recognition on embedded devices.

Please see

• https://github.com/k2-fsa/sherpa-onnx
• https://github.com/k2-fsa/sherpa-ncnn

Note: Please use our latest streaming zipformer for testing.

[fquirin]

I'm certainly interested in adding those engines, especially since I'll probably add 'onnxruntime' with the next update anyways.

Would you say the Sherpa design and API have matured enough now to be considered "ready for production"? What I want to avoid is implementing a very experimental build that still has a lot of moving parts.

[csukuangfj]

sherpa-ncnn can be considered stable for now.

sherpa-onnx is still under development.