README.md

metadata

license: apache-2.0
metrics:
  - cer
base_model:
  - openai/whisper-large-v3

Welcome

If you find this model helpful, please like this model and star us on https://github.com/LianjiaTech/BELLE and https://github.com/shuaijiang/Whisper-Finetune

Belle-whisper-large-v3-zh

Fine tune whisper-large-v3 to enhance Chinese speech recognition capabilities, Belle-whisper-large-v3-zh demonstrates a 24-65% relative improvement in performance on Chinese ASR benchmarks, including AISHELL1, AISHELL2, WENETSPEECH, and HKUST.

Usage

from transformers import pipeline

transcriber = pipeline(
  "automatic-speech-recognition", 
  model="BELLE-2/Belle-whisper-large-v3-zh"
)

transcriber.model.config.forced_decoder_ids = (
  transcriber.tokenizer.get_decoder_prompt_ids(
    language="zh", 
    task="transcribe"
  )
)

transcription = transcriber("my_audio.wav") 

Fine-tuning

Model	(Re)Sample Rate	Train Datasets	Fine-tuning (full or peft)
Belle-whisper-large-v3-zh	16KHz	AISHELL-1 AISHELL-2 WenetSpeech HKUST	full fine-tuning

If you want to fine-thuning the model on your datasets, please reference to the github repo

CER(%) ↓

Model	Language Tag	aishell_1_test(↓)	aishell_2_test(↓)	wenetspeech_net(↓)	wenetspeech_meeting(↓)	HKUST_dev(↓)
whisper-large-v3	Chinese	8.085	5.475	11.72	20.15	28.597
Belle-whisper-large-v2-zh	Chinese	2.549	3.746	8.503	14.598	16.289
Belle-whisper-large-v3-zh	Chinese	2.781	3.786	8.865	11.246	16.440

It is worth mentioning that compared to Belle-whisper-large-v2-zh, Belle-whisper-large-v3-zh has a significant improvement in complex acoustic scenes(such as wenetspeech_meeting).

Citation

Please cite our paper and github when using our code, data or model.

@misc{BELLE,
  author = {BELLEGroup},
  title = {BELLE: Be Everyone's Large Language model Engine},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/LianjiaTech/BELLE}},
}