Reproduce “Token-Shuffle: Towards High-Resolution Image Generation with Autoregressive Models” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xu Ma, Peize Sun, Haoyu Ma, Hao Tang, Chih-Yao Ma, Jialiang Wang, Kunpeng Li, Xiaoliang Dai and 17 more
Token-Shuffle enables autoregressive text-to-image generation at 2048×2048 resolution by grouping nearby visual tokens before Transformer processing and restoring them afterward. With a 2.7B-parameter model, it processes fewer Transformer positions while retaining the final image-token grid. On hard prompts, its VQAScore is 0.77 versus LlamaGen’s 0.59, using length-matched rewritten prompts; efficiency evidence is token counts and FLOPs, not measured runtime.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.