vLLM + Voxtral Setup Issue
#26
by
susovantelecmi
- opened
- Could you please clarify how to use vLLM with Voxtral? I’ve been trying to set it up, but I keep encountering errors even though my system has 80GB of VRAM. If you have any tested script or reference implementation, that would be very helpful.
- I’m trying to transcribe an audio file containing a mix of Hindi and English, but the model keeps translating the English words into Hindi, even though I explicitly instructed it not to translate, and provided an example in the prompt.