TY  - RPRT
TI  - An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation
AU  - Haoran Wang
AU  - Jinchuan Tian
AU  - Siddhant Arora
AU  - Shinji Watanabe
PY  - 2026
UR  - https://arxiv.org/abs/2607.02119
ID  - 2607.02119
ER  -