[Model] Add opt-in HF-numerics mode for the Qwen3-Omni thinker - #8186

[Model] Add opt-in HF-numerics mode for the Qwen3-Omni thinker#8186

amy-why-3459 merged 9 commits into

vllm-project:mainvllm-project/vllm-omni:mainfrom

cr-gao:feat/thinker-hf-numericscr-gao/vllm-omni:feat/thinker-hf-numericsCopy head branch name to clipboard