ProReflow：基於分解速度的漸進式重排

ProReflow: Progressive Reflow with Decomposed Velocity

March 5, 2025

作者: Lei Ke, Haohang Xu, Xuefei Ning, Yu Li, Jiajun Li, Haoling Li, Yuxuan Lin, Dongsheng Jiang, Yujiu Yang, Linfeng Zhang

cs.AI

摘要

擴散模型在圖像和視頻生成領域取得了顯著進展，但仍面臨巨大的計算成本問題。作為一種有效的解決方案，流匹配旨在將擴散模型的擴散過程重新調整為直線，以實現少步甚至一步生成。然而，本文認為流匹配的原始訓練流程並非最優，並引入了兩種技術來改進它。首先，我們提出了漸進式重流，它在局部時間步中逐步重流擴散模型，直至整個擴散過程完成，從而降低了流匹配的難度。其次，我們引入了對齊的v預測，強調了在流匹配中方向匹配的重要性，而非幅度匹配。在SDv1.5和SDXL上的實驗結果證明了我們方法的有效性，例如，在SDv1.5上進行實驗，僅用4個採樣步驟就在MSCOCO2014驗證集上達到了10.70的FID，接近我們的教師模型（32個DDIM步驟，FID = 10.05）。

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).