Chenye Meng
YOU?
Author Swipe
View article: Inversion-DPO: Precise and Efficient Post-Training for Diffusion Models
Inversion-DPO: Precise and Efficient Post-Training for Diffusion Models Open
Recent advancements in diffusion models (DMs) have been propelled by alignment methods that post-train models to better conform to human preferences. However, these approaches typically require computation-intensive training of a base mode…
View article: LAION-SG: An Enhanced Large-Scale Dataset for Training Complex Image-Text Models with Structural Annotations
LAION-SG: An Enhanced Large-Scale Dataset for Training Complex Image-Text Models with Structural Annotations Open
Recent advances in text-to-image (T2I) generation have shown remarkable success in producing high-quality images from text. However, existing T2I models show decayed performance in compositional image generation involving multiple objects …
View article: Distilling Diffusion Models to Efficient 3D LiDAR Scene Completion
Distilling Diffusion Models to Efficient 3D LiDAR Scene Completion Open
Diffusion models have been applied to 3D LiDAR scene completion due to their strong training stability and high completion quality. However, the slow sampling speed limits the practical application of diffusion-based scene completion model…
View article: Distribution Backtracking Builds A Faster Convergence Trajectory for Diffusion Distillation
Distribution Backtracking Builds A Faster Convergence Trajectory for Diffusion Distillation Open
Accelerating the sampling speed of diffusion models remains a significant challenge. Recent score distillation methods distill a heavy teacher model into a student generator to achieve one-step generation, which is optimized by calculating…
View article: Learning Object Consistency and Interaction in Image Generation from Scene Graphs
Learning Object Consistency and Interaction in Image Generation from Scene Graphs Open
This paper is concerned with synthesizing images conditioned on a scene graph (SG), a set of object nodes and their edges of interactive relations. We divide existing works into image-oriented and code-oriented methods. In our analysis, th…