vllm

31a08f5b - [Model] Add min_pixels / max_pixels to Qwen2VL as mm_processor_kwargs (#9612)