Multi-Scale Local Speculative Decoding for Image Generation 事件

Multi-Scale Local Speculative Decoding for Image Generation arXiv:2601.05149v2 Announce Type: replace Abstract: Autoregressive (AR) models have achieved remarkable success in image synthesis, yet their sequential nature imposes significant latency constraints. Speculative Decoding offers a promising avenue for acceleration, but existing approaches are limited by token-level ambiguity and lack of spatial awareness. In this work, we introduce Multi-Scale Local Speculative Decoding (MuLo-SD), a no