search

huggingface / distil-whisper

Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.
MIT License
3.33k stars 238 forks source link

How can I assign parameters for another language transcribe task? #62

Closed liuamo closed 5 months ago

liuamo commented 6 months ago

user this methods can not resolved problem. model_large = hf_hub_download(model_id, filename="original-model.bin") model = load_model(model_large) result = model.transcribe("./audio.mp3", language='Chinese')

sanchit-gandhi commented 6 months ago

Hey @liuamo - as mentioned on the README, the 'official' checkpoints are for English-only: https://github.com/huggingface/distil-whisper#distil-whisper

But you can quite easily distil a Whisper model of your own for Mandarin using the provided training code: https://github.com/huggingface/distil-whisper/tree/main/training

liuamo commented 6 months ago

Thanks for your reply!

Hey @liuamo - as mentioned on the README, the 'official' checkpoints are for English-only: https://github.com/huggingface/distil-whisper#distil-whisper

But you can quite easily distil a Whisper model of your own for Mandarin using the provided training code: https://github.com/huggingface/distil-whisper/tree/main/training