4 papers
DART: Distillation-Aware Reparameterization for Training-Free LoRA Reuse in Few-Step Video Diffusion Models
Shihong Li, Juntao Xu, JinCao +3
Few-step distillation reduces the inference cost of image-to-video generation, but directly reusing LoRA adapters trained for long denoising trajectories can weaken their intended…
ToPO: Token-Oriented Preference Optimization for Text-to-Image Diffusion Models
Juntao Xu, Shihong Li, Hoi Fan Au +1
Diffusion models have achieved remarkable success in text-to-image generation, yet aligning their outputs with human preferences remains challenging. Diffusion-DPO learns from imag…
One Knob to Rule Them All: A Unified Optimal Transport View of Cold-Start Active Learning
Ning Zhu, Xiaochuan Ma, Juntao Xu +4
Cold-Start Active Learning (CSAL) aims to select a valuable subset from an unlabeled pool without any prior knowledge or human assistance. Existing methods take diverse routes base…
Rectify Then Diffuse: Disentangling Concepts Before Denoising Trajectory Unfolds
Ning Zhu, An Chen, Mengfei Zhao +4
Text-to-image diffusion models can generate individual concepts well, but they often omit or merge concepts incorrectly with multiple concepts. We trace these failures to an early…