Availability of alternative translation models and speech recognition models

[wqwwffsytxwp]

For example, Alibaba's SenseVoice is very good at recognizing in certain scenarios
https://github.com/FunAudioLLM/SenseVoice

[niedev]

Hi, thank you for your suggestion, I am already working on that (now I am improving the downloader), soon I will support the choose of 3 models for translation: NLLB Q4, NLLB Q8 (used now) and Madlad Q4. In the future I will also support different size of Whisper, and I will evaluate the quality of SenseVoice, seems interesting, stay tuned 🚀