用连续流方法加速类别数据生成，支持单步高质量采样。
Categorical Flow Maps
- 基于连续流匹配与自蒸馏，构建约束概率分布的生成路径。
- 在图像、分子图和文本上实现少步生成最优性能，单步也表现优异。
- 兼容现有引导技术，适合需快速生成的下游任务场景。
我们提出类别流映射（Categorical Flow Maps），一种通过自蒸馏实现类别数据快速少步生成的流匹配方法。基于近期变分流匹配框架及扩散与流模型加速推理的趋势，我们定义了一种指向单纯形的流映射，将概率质量朝向预测终点迁移，从而自然约束模型输出。由于轨迹为连续而非离散，该方法可使用现有蒸馏技术，也可采用基于终点一致性的新目标进行训练。这一连续形式还天然支持测试时推理：可直接复用现有引导与重加权技术，在类别设置中引导采样以达成下游目标。实验表明，在图像、分子图和文本任务上，该方法实现了当前最优的少步生成效果，甚至在单步生成中也表现出色。
原文摘要 · Abstract (English)
We introduce Categorical Flow Maps, a flow-matching method for accelerated few-step generation of categorical data via self-distillation. Building on recent variational formulations of flow matching and the broader trend towards accelerated inference in diffusion and flow-based models, we define a flow map towards the simplex that transports probability mass toward a predicted endpoint, yielding a parametrisation that naturally constrains model predictions. Since our trajectories are continuous rather than discrete, Categorical Flow Maps can be trained with existing distillation techniques, as well as a new objective based on endpoint consistency. This continuous formulation also automatically unlocks test-time inference: we can directly reuse existing guidance and reweighting techniques in the categorical setting to steer sampling toward downstream objectives. Empirically, we achieve state-of-the-art few-step results on images, molecular graphs, and text, with strong performance even in single-step generation.
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。