{"slug": "data-forcing-distillation-restoring-diversity-and-fidelity-in-few-step-video", "title": "Data-Forcing Distillation: Restoring Diversity and Fidelity in Few-Step Video Generation", "summary": "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.", "body_md": "arXiv:2606.18478v1 Announce Type: new\nAbstract: 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.", "url": "https://wpnews.pro/news/data-forcing-distillation-restoring-diversity-and-fidelity-in-few-step-video", "canonical_source": "https://arxiv.org/abs/2606.18478", "published_at": "2026-06-18 04:00:00+00:00", "updated_at": "2026-06-19 02:00:51.431217+00:00", "lang": "en", "topics": ["machine-learning", "generative-ai", "computer-vision", "ai-research", "large-language-models"], "entities": ["Wan2.1-1.3B", "Cosmos-Predict2.5-2B", "Distribution Matching Distillation", "DMD2", "Data-Forcing Distillation"], "alternates": {"html": "https://wpnews.pro/news/data-forcing-distillation-restoring-diversity-and-fidelity-in-few-step-video", "markdown": "https://wpnews.pro/news/data-forcing-distillation-restoring-diversity-and-fidelity-in-few-step-video.md", "text": "https://wpnews.pro/news/data-forcing-distillation-restoring-diversity-and-fidelity-in-few-step-video.txt", "jsonld": "https://wpnews.pro/news/data-forcing-distillation-restoring-diversity-and-fidelity-in-few-step-video.jsonld"}}