提出SFTok，让离散图像分词器在高压缩下实现顶尖重建质量。
SFTok: Bridging the Performance Gap in Discrete Tokenizers
- 采用多步迭代与自强化引导重建，解决训练-推理不一致问题。
- 64个令牌压缩率下，ImageNet上重建质量（rFID=1.21）达顶尖水平。
- 适合追求高效、高质量图像生成的多模态系统开发者。
近期多模态模型的发展凸显了图像分词在高分辨率图像生成中的关键作用。通过将图像压缩为紧凑的潜在表示，分词器使生成模型能在低维空间运行，从而提升计算效率并降低复杂度。尽管离散分词器天然契合自回归范式，但其性能仍落后于连续分词器，限制了在多模态系统中的应用。为此，我们提出SFTok，一种结合多步迭代机制的离散分词器，通过引入自强化引导视觉重建和去偏-拟合训练策略，有效解决了多步过程中的训练-推理不一致问题，显著提升了图像重建质量。在仅使用64个令牌每图像的高压缩率下，SFTok在ImageNet上达到rFID=1.21的先进重建质量，并在类到图像生成任务中表现优异（gFID=2.29）。
原文摘要 · Abstract (English)
Recent advances in multimodal models highlight the pivotal role of image tokenization in high-resolution image generation. By compressing images into compact latent representations, tokenizers enable generative models to operate in lower-dimensional spaces, thereby improving computational efficiency and reducing complexity. Discrete tokenizers naturally align with the autoregressive paradigm but still lag behind continuous ones, limiting their adoption in multimodal systems. To address this, we propose \textbf{SFTok}, a discrete tokenizer that incorporates a multi-step iterative mechanism for precise reconstruction. By integrating \textbf{self-forcing guided visual reconstruction} and \textbf{debias-and-fitting training strategy}, SFTok resolves the training-inference inconsistency in multi-step process, significantly enhancing image reconstruction quality. At a high compression rate of only 64 tokens per image, SFTok achieves state-of-the-art reconstruction quality on ImageNet (rFID = 1.21) and demonstrates exceptional performance in class-to-image generation tasks (gFID = 2.29).
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。