Toward Parallel Serving for Vision-Language Models via Modal Decoupling and Scheduling.

Yijia Yang, Yubo Deng, Yida Wang, Yuanchao Xu, Keni Qiu

Browse the full DATE paper archive.