sherpa-onnx
sherpa-onnx copied to clipboard
k2-fsa
ASR English model cannot recognize my voice well
why all the english related model in asr cannot recognize well? though Chinese ones seem work well.
Have you tested the english related models? Am i missing anything to make it accurate? thanks.
e.g. this model:sherpa-ncnn-streaming-zipformer-20m-2023-02-17
I don't think it is about my english tone, because even i use the standard english audio, it doesn't work well. the recognized result seem bad.
i have test all the models listed sherpa, i don't think any of them is good enough to recognize english, though chinese one is good
btw, please see : https://k2-fsa.github.io/icefall/model-export/export-ncnn-conv-emformer.html#export-the-model-via-torch-jit-trace
i have test the model in both onnx and ncnn examples for swift, they cannot recognize English well, i can say it is very bad, WER is at least 50%, am i missing anything? can you try the model using the swift sample for a try? thanks.
【新一代 Kaldi: 最新版 zipformer 在 iOS 上的英文语音识别演示-哔哩哔哩】 https://b23.tv/K87JkfW
【新一代 Kaldi: Two-pass 实时英语语音识别之 iOS （奥巴马演讲)-哔哩哔哩】 https://b23.tv/YS2QNmR
We have already provided Engine asr video demos
If it is not working for you, please recheck.
@csukuangfj I have got same error. When run flutter streaming_asr example using english model like sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2.
they cannot recognize English when stop and start record, watch video demo error Watch on youtube
I try sherpa-onnx-streaming-zipformer-en-2023-06-21. seem like it work. But problem is this model too large for mobile. I try Chinese language with small model asr-models/sherpa-onnx-streaming-zipformer-zh-14M-2023-02-23 It still working. Can you make English model working with smaller size
Model sherpa-onnx-streaming-zipformer-en-2023-06-21 ~ 500M. So It make App size > 600M. I think App with that size is big.
Model sherpa-onnx-streaming-zipformer-en-2023-06-21 ~ 500M
No, we don't have model files that large.
Please keep only the files you use in the code and remove other unused files.
There is no need to keep unused files in your app.
I suggest that you have a look at the Flutter apps we provide
https://k2-fsa.github.io/sherpa/onnx/flutter/pre-built-app.html#streaming-speech-recognition-stt-asr
Are our APPs that large?
Please keep only the files you use in the code and remove other unused files.
Thank you very much! Now I know how to make App size smaller