alphacep / vosk-api

Offline speech recognition API for Android, iOS, Raspberry Pi and servers with Python, Java, C# and Node
Apache License 2.0
7.57k stars 1.06k forks source link

Multilingual recognition #1326

Closed MehdiyevSoltan2004 closed 1 year ago

MehdiyevSoltan2004 commented 1 year ago

How can I use vosk for multilingual language recognition? Are there ready-to-use multilingual models? Or is there any way to supplement one model with another language? Please answer here and in detail, without referring to the other discussion, because nothing is so clear

nshmyrev commented 1 year ago

How can I use vosk for multilingual language recognition?

Load multilingual model.

Are there ready-to-use multilingual models?

Our Chinese model recognizes English too. We also have English-Hindi model, not published yet though.

MehdiyevSoltan2004 commented 1 year ago

Is there any way I can use .pt files as Vosk models?

nshmyrev commented 1 year ago

No, usually pt models have their own toolkits

MehdiyevSoltan2004 commented 1 year ago

Do you have models that support English and Russian? Or can I use several models at the same time?

nshmyrev commented 1 year ago

We do not have English + Russian model

MehdiyevSoltan2004 commented 1 year ago

Ok, thanks

nshmyrev commented 1 year ago

Same as #233

I8Robot commented 1 year ago

How can I use vosk for multilingual language recognition?

Load multilingual model.

Are there ready-to-use multilingual models?

Our Chinese model recognizes English too. We also have English-Hindi model, not published yet though.

Where can I get the Chinese model which can recognizes English. Is it published?

shivashankar6146 commented 11 months ago

How can I use vosk for multilingual language recognition?

Load multilingual model.

Are there ready-to-use multilingual models?

Our Chinese model recognizes English too. We also have English-Hindi model, not published yet though.

@nshmyrev Do you have plans to publish the English-Hindi model anytime soon?

nshmyrev commented 11 months ago

@shivashankar6146 we plan it but there is no good stimulus unfortunately

nshmyrev commented 11 months ago

@I8Robot yes, standard https://alphacephei.com/vosk/models/vosk-model-cn-0.22.zip should recognize English words. You can also check

https://k2-fsa.github.io/sherpa/onnx/pretrained_models/online-transducer/zipformer-transducer-models.html#csukuangfj-sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20-bilingual-chinese-english

or https://k2-fsa.github.io/sherpa/onnx/pretrained_models/online-paraformer/paraformer-models.html#csukuangfj-sherpa-onnx-streaming-paraformer-bilingual-zh-en-chinese-english

paraformer is very good for Chinese

shivashankar6146 commented 11 months ago

Thanks for the update👍 Will keep watching this space for further updates regarding this model esply.