SII-Auraithm
SII-Auraithm
### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...
## Motivation In DLLM (Disaggregated LLM) mode, tokens are generated in blocks and progressively unmasked in a non-sequential order. Currently, there is no way to track which decoding step each...
### Checklist - [ ] 1. I have searched related issues but cannot get the expected help. - [ ] 2. The bug has not been fixed in the latest...
### Checklist - [ ] 1. I have searched related issues but cannot get the expected help. - [ ] 2. The bug has not been fixed in the latest...
### Your current environment vllm==0.10.0 ### 🐛 Describe the bug I tried to pass the parameter use_audio_in_video. ``` from openai import OpenAI API_KEY = "12" client = OpenAI( api_key=API_KEY, base_url="http://0.0.0.0:8091/v1",...
### Your current environment vllm==0.12.0 vllm-omni==0.12.0rc1 ### 🐛 Describe the bug Command ``` VLLM_SERVER_DEV_MODE=1 vllm serve /xxx/Qwen3-Omni-30B-A3B-Instruct --omni --port 8901 --gpu_memory_utilization 0.9 -tp 8 I faced the error ``` ERROR...