8 papers
DualHeads: Reward-tilted few-step diffusion made simple
Yuanzhi Zhu, Eleftherios Tsonis, Simo Alami Chehboune +3
Distilled diffusion models generate high-quality images in a few sampling steps, but aligning them with task requirements and human preferences is difficult. Methods developed for…
MFSR: MeanFlow Distillation for One Step Real-World Image Super Resolution
Ruiqing Wang, Kai Zhang, Yuanzhi Zhu +3
Diffusion- and flow-based models have advanced Real-world Image Super-Resolution (Real-ISR), but their multi-step sampling makes inference slow and hard to deploy. One-step distill…
What about gravity in video generation? Post-Training Newton's Laws with Verifiable Rewards
Minh-Quan Le, Yuanzhi Zhu, Vicky Kalogeiton +1
Recent video diffusion models can synthesize visually compelling clips, yet often violate basic physical laws-objects float, accelerations drift, and collisions behave inconsistent…
One-step Diffusion Models with Bregman Density Ratio Matching
Yuanzhi Zhu, Eleftherios Tsonis, Lucas Degeorge +1
Diffusion and flow models achieve high generative quality but remain computationally expensive due to slow multi-step sampling. Distillation methods accelerate them by training fas…
Soft-Di[M]O: Improving One-Step Discrete Image Generation with Soft Embeddings
Yuanzhi Zhu, Xi Wang, Stéphane Lathuilière +1
One-step generators distilled from Masked Diffusion Models (MDMs) compress multiple sampling steps into a single forward pass, enabling efficient text and image synthesis. However,…
DiO: Distilling Masked Diffusion Models into One-step Generator
Yuanzhi Zhu, Xi Wang, Stéphane Lathuilière +1
Masked Diffusion Models (MDMs) have emerged as a powerful generative modeling technique. Despite their remarkable results, they typically suffer from slow inference with several st…
Bridging Continuous and Discrete Tokens for Autoregressive Visual Generation
Yuqing Wang, Zhijie Lin, Yao Teng +4
Autoregressive visual generation models typically rely on tokenizers to compress images into tokens that can be predicted sequentially. A fundamental dilemma exists in token repres…
SlimFlow: Training Smaller One-Step Diffusion Models with Rectified Flow
Yuanzhi Zhu, Xingchao Liu, Qiang Liu
Diffusion models excel in high-quality generation but suffer from slow inference due to iterative sampling. While recent methods have successfully transformed diffusion models into…