VA-TDM boosts few-step diffusion models
Original title: VA-TDM: A Unified View for Improving Few-Step Diffusion Models by Velocity Adversarial Distillation
VA-TDM uses velocity adversarial distillation to reach state-of-the-art four-step text-to-image and text-to-video generation.
Few-step distillation for diffusion and flow models is challenging due to distribution tracking issues. VA-TDM introduces velocity adversarial distillation with a learned critic to unify score, Fisher, and moment matching. It delivers state-of-the-art results across backbones and scales to 33B MiniMax-H3.
- State-of-the-art four-step performance
- Combines trajectory distribution matching
- Outperforms Fast-H3 on MiniMax-H3
- Effective for text-to-video
- Organizations
- Hugging Face
From the original article
1 min read at huggingface.co
Few-step distillation offers efficient generation with diffusion and flow models, but preserving quality with only a few sampling steps remains challenging. Distribution-matching methods via score distillation provide a strong foundation, yet they rely on accurately tracking the student's changing distribution.
By Yihong Luo, Tianyu Pang, Tianyang Hu · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.