Open
Description
🚀 The feature, motivation and pitch
Does VLLM currently support deploying the Qwen/Qwen3-235B-A22B-FP8 model using the PD separation + DP + EP + DeepEP method?
Alternatives
No response
Additional context
No response
Before submitting a new issue...
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.