locality-aware-parallel-autoregressive-decoding

【效率与协作】Accelerate image generation by generating multiple patches in parallel instead of sequentially. Uses locality-aware ordering that generates spatially-close tokens while keeping concurrent tokens far apart, reducing steps from 256 to 20 (12× speedup) while maintaining quality.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准