vLLM Just Merged TurboQuant Fix for Qwen 3.5+

Posted by havenoammo@reddit | LocalLLaMA | View on Reddit | 27 comments

Previously it was throwing a 'Not Implemented' error due to Mamba layers. Going to test it now!

https://github.com/vllm-project/vllm/pull/39931