CaTok：驯服平均流实现一维因果图像分词
CaTok: Taming Mean Flows for One-Dimensional Causal Image Tokenization
把图像压缩成一维因果词元序列并用 MeanFlow 解码器做一步/多步采样，可直接作为视觉自回归生成的前端分词器替换方案；REPA-A 用视觉基础模型特征对齐编码器，是一种可复用的稳定训练正则化手段。
原文依据 we present CaTok, a 1D causal image tokenizer with a MeanFlow decoder
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论自回归语言模型依赖因果分词，但将该范式扩展到视觉领域仍非易事。现有的视觉分词器要么将二维图块展平为非因果序列，要么施加与「下一词元预测」模式相错位的启发式排序。近期的扩散自编码器同样存在不足：让解码器以全部词元为条件缺乏因果性，而采用嵌套 dropout 机制又会引入不平衡。为应对这些挑战，本文提出 CaTok，一种带 MeanFlow 解码器的一维因果图像分词器。通过按时间区间选取词元并将其与 MeanFlow 目标绑定，CaTok 学习到因果的一维表征，既支持快速的一步生成，也支持高保真的多步采样，同时在不同词元区间上自然捕捉多样的视觉概念。为进一步稳定并加速训练，本文提出一种简洁的正则化方法 REPA-A，将编码器特征与视觉基础模型对齐。实验表明，CaTok 在 ImageNet 重建上取得最优结果，以更少的训练轮数达到 0.75 FID、22.53 峰值信噪比和 0.674 结构相似度，其自回归模型也取得与领先方法相当的性能。项目网站见 https://sharelab-sii.github.io/catok-web。
英文原文摘要 1318 字符
Autoregressive (AR) language models rely on causal tokenization, but extending this paradigm to vision remains non-trivial. Current visual tokenizers either flatten 2D patches into non-causal sequences or enforce heuristic orderings that misalign with the "next-token prediction" pattern. Recent diffusion autoencoders similarly fall short: conditioning the decoder on all tokens lacks causality, while applying nested dropout mechanism introduces imbalance. To address these challenges, we present CaTok, a 1D causal image tokenizer with a MeanFlow decoder. By selecting tokens over time intervals and binding them to the MeanFlow objective, CaTok learns causal 1D representations that support both fast one-step generation and high-fidelity multi-step sampling, while naturally capturing diverse visual concepts across token intervals. To further stabilize and accelerate training, we propose a straightforward regularization REPA-A, which aligns encoder features with Vision Foundation Models (VFMs). Experiments demonstrate that CaTok achieves state-of-the-art results on ImageNet reconstruction, reaching 0.75 FID, 22.53 PSNR and 0.674 SSIM with fewer training epochs, and the AR model attains performance comparable to leading approaches. Project website is available in https://sharelab-sii.github.io/catok-web.