随机自回归视觉生成
Randomized Autoregressive Visual Generation
AI总结：
本研究提出随机自回归建模RAR，通过训练中随机置换输入序列顺序并退火的策略，在保留自回归框架与语言建模兼容性的同时，在ImageNet-256上取得1.48的FID，刷新图像生成SOTA。
摘要：
本文提出了用于视觉生成的随机自回归建模（Randomized AutoRegressive modeling，RAR）方法，该方法在图像生成任务上取得了全新的最优性能，同时与语言建模框架完全兼容。提出的RAR方法原理简单：在以预测下一个token为目标的标准自回归训练过程中，通常以光栅形式排列的输入序列会以概率r被随机置换为不同的分解顺序，其中r初始值为1，并在训练过程中线性衰减至0。这种退火训练策略使模型能够学习最大化所有分解顺序下的期望似然，从而有效提升模型对双向上下文的建模能力。重要的是，RAR保留了自回归建模框架的完整性，确保与语言建模完全兼容，同时显著提升了图像生成性能。在ImageNet-256基准测试上，RAR取得了1.48的FID分数，不仅超越了此前最优的自回归图像生成器，还优于领先的基于扩散和基于掩码Transformer的方法。代码和模型将在https://github.com/bytedance/1d-tokenizer 发布。
英文摘要：
This paper presents Randomized AutoRegressive modeling (RAR) for visual generation, which sets a new state-of-the-art performance on the image generation task while maintaining full compatibility with language modeling frameworks. The proposed RAR is simple: during a standard autoregressive training process with a next-token prediction objective, the input sequence-typically ordered in raster form-is randomly permuted into different factorization orders with a probability r, where r starts at 1 and linearly decays to 0 over the course of training. This annealing training strategy enables the model to learn to maximize the expected likelihood over all factorization orders and thus effectively improve the model's capability of modeling bidirectional contexts. Importantly, RAR preserves the integrity of the autoregressive modeling framework, ensuring full compatibility with language modeling while significantly improving performance in image generation. On the ImageNet-256 benchmark, RAR achieves an FID score of 1.48, not only surpassing prior state-of-the-art autoregressive image generators but also outperforming leading diffusion-based and masked transformer-based methods. Code and models will be made available at https://github.com/bytedance/1d-tokenizer
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
