作者位次 #7/10ACL 2026Poster· Natural Language Processing
Uni- MMMU : A Massive Multi-discipline Multimodal Unified Benchmark
Kai Zou · Ziqi Huang · Yuhao Dong · Shulin Tian · Dian Zheng · Hongbo Liu · Jingwen He · Bin Liu · Yu Qiao · Ziwei Liu
摘要
Unified multimodal models aim to jointly enable visual understanding and generation, yet current benchmarks rarely examine their true integration. Existing evaluations either treat the two abilities in isolation or overlook tasks that inherently couple them. To address this gap, we present Uni-MM...