k2-fsa / sherpa-ncnn

Real-time speech recognition and voice activity detection (VAD) using next-gen Kaldi with ncnn without Internet connection. Support iOS, Android, Linux, macOS, Windows, Raspberry Pi, VisionFive2, LicheePi4A etc.
https://k2-fsa.github.io/sherpa/ncnn/index.html
Apache License 2.0
968 stars 146 forks source link

[Help wanted] Add the small streaming zipformer model from icefall #903 #107

Open csukuangfj opened 1 year ago

csukuangfj commented 1 year ago

https://github.com/k2-fsa/icefall/pull/903 provides a smaller streaming zipformer.

TODOs

marcoyang1998 commented 1 year ago

I will do this.

marcoyang1998 commented 1 year ago

The converted model can be found here https://huggingface.co/marcoyang/sherpa-ncnn-streaming-zipformer-20M-2023-02-17

joazoa commented 1 year ago

Is the new Improved Zipformer already supported in ncnn ?

csukuangfj commented 1 year ago

Is the new Improved Zipformer already supported in ncnn ?

No, not yet. We have not converted it to ncnn format yet.

joazoa commented 6 months ago

@csukuangfj do you know if anybody is working on wasm support for onnx or ncnn support for latest zipformer ? (So any way to run the latest zipformer with wasm)

csukuangfj commented 6 months ago

@csukuangfj do you know if anybody is working on wasm support for onnx or ncnn support for latest zipformer ? (So any way to run the latest zipformer with wasm)

I am working on the wasm support of the latest zipformer in sherpa-onnx.

By the way, we have already supported running tts with wasm in sherpa-onnx, which means we have already setup the environment for wasm development and it should not be difficult to support asr with wasm in sherpa-onnx.

csukuangfj commented 6 months ago

@csukuangfj do you know if anybody is working on wasm support for onnx or ncnn support for latest zipformer ? (So any way to run the latest zipformer with wasm)

It has been supported in sherpa-onnx. We support running it both in the browser and in NodeJS via WebAssembly. @joazoa

joazoa commented 6 months ago

Thank you!!