Missing vllm Integration for Accelerated Inference in This Repository

[zhaochen0110]

Hiiii,

Great job!! But I noticed that OpenR1 supports accelerated inference using vllm, which can significantly improve performance. However, I couldn't find any integration of this feature in this repository.

Is this an intentional omission, or have I possibly overlooked some documentation or configuration option that enables vllm support?

[baibizhe]

yes,This repo does not support vllm.