Fun-ASR-Nano-2512 cantonese performance is not good

[ming030890]

Thanks for the effort! However, I ran some evaluations, and Fun-ASR-Nano-2512 has a higher error rate than SenseVoice and WenetSpeech-Yue in terms of Cantonese performance.

Eval sets:
ming030890/cantonese_asr_eval_mdcc_long
commonvoice
ming030890/youtube_caption_yue

[LauraGPT]

Do you use it with vad model?

[ming030890]

Yes, I have tried both, i.e. with vad and without vad.

Attached the full eval results.
Note that cv means commonvoice.

Thanks!