Distribution Backtracking Builds A Faster Convergence Trajectory for Diffusion Distillation

Documenti analoghi

• Inversion-DPO: Precise and Efficient Post-Training for Diffusion Models
  di: Li, Zejian, et al.
  Pubblicazione: (2025)
• LAION-SG: An Enhanced Large-Scale Dataset for Training Complex Image-Text Models with Structural Annotations
  di: Li, Zejian, et al.
  Pubblicazione: (2024)
• Distilling Diffusion Models to Efficient 3D LiDAR Scene Completion
  di: Zhang, Shengyuan, et al.
  Pubblicazione: (2024)
• Diffusion Distillation With Direct Preference Optimization For Efficient 3D LiDAR Scene Completion
  di: Zhao, An, et al.
  Pubblicazione: (2025)
• Beyond Binary Preference: Aligning Diffusion Models to Fine-grained Criteria by Decoupling Attributes
  di: Meng, Chenye, et al.
  Pubblicazione: (2026)