Saining Xie· @sainingxie · X·· 2026-05-22AI 评分60
AI 导读
RAEv2通过大幅简化架构并提升通用性，在文本到图像（T2I）和世界模型等任务中实现了超过10倍的收敛速度提升，同时改善了重建与生成质量。研究团队在大量实验中发现，强大的表示编码器对像素解码器至关重要。传统评估指标（如FID）已不足以全面衡量模型性能，新的评估指标（如ep@fid-k/fdr^k）揭示了生成模型领域仍存在广阔的研究空间。
正文 · 原文
check out RAEv2 led by Jas. through extensive exps, we found some really intriguing behaviors showing why strong representation encoders are key for pixel decoders.
spoiler: it’s not about hillclimbing fid; new metrics like ep@fid-k/fdr^k show there’s a lot more left to explore!
In Oct last year, Representation Autoencoders provided an elegant solution to unified tokenization for understanding and generation. Today we make them a bit more simple. a bit more general. Result: >10x faster convergence, better reconstruction, better generation. And yes we test them on T2I and world models :) Introducing RAEv2在 X 查看被引用的帖子
来源：Saining Xie · x.com