用对角线顺序生成图像，让相邻像素更接近，提升生成质量。
Direction-Aware Diagonal Autoregressive Image Generation
- 采用对角线扫描顺序生成图像，保持相邻索引的像素邻近。
- 20亿参数模型在ImageNet上达1.37的FID，刷新自回归生成纪录。
- 适合关注图像生成细节与结构一致性的研究者使用。
栅格顺序的图像标记序列在行切换处存在显著的欧氏距离，不利于自回归生成。本文提出方向感知对角线自回归图像生成方法（DAR），按对角线扫描顺序生成图像标记。该顺序确保相邻索引的标记在空间上仍保持紧密，同时允许因果注意力从更广范围获取信息。此外，引入4D-RoPE和方向嵌入两个方向感知模块，增强模型处理生成方向频繁变化的能力。为充分利用图像分词器的表示能力，采用其码本作为图像标记嵌入。提出485M至20亿参数的多个模型。在256×256 ImageNet基准上，DAR-XL（20亿参数）超越所有先前自回归图像生成器，达到1.37的FID分数，达到当前最优水平。
原文摘要 · Abstract (English)
The raster-ordered image token sequence exhibits a significant Euclidean distance between index-adjacent tokens at line breaks, making it unsuitable for autoregressive generation. To address this issue, this paper proposes Direction-Aware Diagonal Autoregressive Image Generation (DAR) method, which generates image tokens following a diagonal scanning order. The proposed diagonal scanning order ensures that tokens with adjacent indices remain in close proximity while enabling causal attention to gather information from a broader range of directions. Additionally, two direction-aware modules: 4D-RoPE and direction embeddings are introduced, enhancing the model's capability to handle frequent changes in generation direction. To leverage the representational capacity of the image tokenizer, we use its codebook as the image token embeddings. We propose models of varying scales, ranging from 485M to 2.0B. On the 256$\times$256 ImageNet benchmark, our DAR-XL (2.0B) outperforms all previous autoregressive image generators, achieving a state-of-the-art FID score of 1.37.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。