Quentin Fuxa

Results 74 comments of Quentin Fuxa

> always busy. - I love it. - I love it. There is so much to see and do. - I like my city as well. - But that is...

Okay thank you for the precision. So (part of) the issue is that you use `wlk --model large-v3-turbo --language zh-CN -l INFO --warmup-file ./jfk.wav` If you look at https://github.com/QuentinFuxa/WhisperLiveKit/blob/main/docs/supported_languages.md ,...

Hi, yeah the model struggles when it does not have enough tokens to work on (beginning of sentences). I plan to do a fix on that

Work in progress, will be in the next release

Hi, a fix for that has been made in 0.2.11.post1 version

Hi, that is probably unrelated, I will look at it

You can develop a frontend in any language, as long as you use the websocket /asr endpoint. It is not in the project's plans to develop a native frontend for...

Yeah, I dont have several gpus to test, so I am not sure

Hi, this error means your system can’t find a compatible cuDNN library. Faster-Whisper requires that your CUDA + cuDNN versions match the PyTorch/CTranslate2 builds you've installed. On Arch/CachyOS the easiest...

Hi, yess having a stable (and faster) translation is work in progress, will be in the next release