通过分阶段重流与方向对齐，4步生成媲美32步扩散模型。
ProReflow: Progressive Reflow with Decomposed Velocity
- 分阶段局部重流降低流匹配难度
- 4步采样在MSCOCO上达FID 10.70，接近32步教师模型
- 适合追求高效生成的图像/视频生成研究者
扩散模型在图像和视频生成中取得显著进展，但仍面临巨大计算开销。流匹配作为有效解决方案，将扩散过程重流为直线，实现少步甚至单步生成。然而本文指出，原始流匹配训练流程并非最优，并提出两项改进：首先引入分阶段重流，逐步在局部时间步重流扩散过程，直至整体完成，降低流匹配难度；其次提出对齐v预测，强调方向匹配优于幅度匹配的重要性。在SDv1.5和SDXL上的实验表明，该方法有效，例如在SDv1.5上仅用4步采样即在MSCOCO2014验证集达到FID 10.70，接近教师模型（32步DDIM，FID = 10.05）。
原文摘要 · Abstract (English)
Diffusion models have achieved significant progress in both image and video generation while still suffering from huge computation costs. As an effective solution, flow matching aims to reflow the diffusion process of diffusion models into a straight line for a few-step and even one-step generation. However, in this paper, we suggest that the original training pipeline of flow matching is not optimal and introduce two techniques to improve it. Firstly, we introduce progressive reflow, which progressively reflows the diffusion models in local timesteps until the whole diffusion progresses, reducing the difficulty of flow matching. Second, we introduce aligned v-prediction, which highlights the importance of direction matching in flow matching over magnitude matching. Experimental results on SDv1.5 and SDXL demonstrate the effectiveness of our method, for example, conducting on SDv1.5 achieves an FID of 10.70 on MSCOCO2014 validation set with only 4 sampling steps, close to our teacher model (32 DDIM steps, FID = 10.05).
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。