1 paper
Yichen Zhang, Da Peng, Zonghao Guo +19
A recent cutting-edge topic in multimodal modeling is to unify visual comprehension and generation within a single model. However, the two tasks demand mismatched decoding regimes…