Is real-time "Speech-to-Speech translation" possible? #603

SlavikCA started this conversation in General

SlavikCA
Nov 26, 2025

qwen3omni has this demo:
https://github.com/QwenLM/Qwen3-Omni/blob/main/cookbooks/speech_translation.ipynb

but lt's not realtime (file used),
and looks like it using custom VLLM build.

Is it possible to run mlx-vlm for real-time "Speech-to-Speech translation" scenario?

Replies: 0 comments

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment