IAR2: Improving Autoregressive Visual Generation with Semantic-Detail Associated Token Prediction

ArXi:2510.06928v2 Announce Type: replace Autoregressive models have emerged as a powerful paradigm for visual content creation, but often overlook the intrinsic structural properties of visual data. Our prior work, IAR, initiated a direction to address this by reorganizing the visual codebook based on embedding similarity, thereby improving generation robustness. However, it is constrained by the rigidity of pre-trained codebooks and the inaccuracies of hard, uniform clustering.