Data-Forcing Distillation: Restoring Diversity and Fidelity in Few-Step Video Generation

Researchers propose Data-Forcing Distillation (DFD), a post-training framework that restores diversity and fidelity in few-step video generation models by addressing mode collapse and over-saturation artifacts. DFD achieves this with minimal code changes and outperforms teacher models on Wan2.1-1.3B and Cosmos-Predict2.5-2B.

arXiv:2606.18478v1 Announce Type: new Abstract: Recent progress has shown promise in distilling multi-step video diffusion models into efficient few-step students. Among them, Distribution Matching Distillation DMD and its successor DMD2 achieved strong generation quality and fast convergence. However, due to the nature of the reverse Kullback--Leibler KL objective, these methods exhibit two persistent failure modes: a substantial drop in sample diversity, and visibly over-saturated outputs that deviate from real-video appearance. In this work, we propose Data-Forcing Distillation DFD , a simple post-training framework that restores diversity and fidelity in DMD with only a single-line of code change. At its core is the teacher score discrepancy to guide the student toward the real-data distribution, pulling it to missing modes mitigating mode collapse and away from problematic modes absent in real data avoiding over-saturation . We provide an in-depth theoretical analysis of our framework and validate our approach on text-to-video, image-to-video, and autoregressive video generation. With only 100--300 steps of finetuning, DFD effectively restores diversity and fidelity on both Wan2.1-1.3B and Cosmos-Predict2.5-2B model, resolving the over-saturation artifacts with significantly better video dynamics and appearance, and even outperforms the teacher model.