文章#

• [2025.11] RubricRL: Simple Generalizable Rewards for Text-to-Image Generation 模型自动生成评分规则
• [2025.10] The Principles of Diffusion Models
  • 扩散模型的基本原理
• [2025.06] An Introduction to Flow Matching and Diffusion Models
• [2025.06] Orthogonal Finetuning Made Scalable oftv2 性能和效果跟lora差不多
• [2025.04] In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer
• [2023.05] Training Diffusion Models with Reinforcement Learning 用VLM+RL训练Diffusion模型
  • jannerm/ddpo
• [2025.05] DanceGRPO: Unleashing GRPO on Visual Generation
• [2023.06] Controlling Text-to-Image Diffusion by Orthogonal Finetuning oft 性能比lora差
• [2022.10] Flow Matching for Generative Modeling
• [2025.03] Wan: Open and Advanced Large-Scale Video Generative Models
  • Wan-Video/Wan2.2
• [2025.03] Open-Sora 2.0: Training a Commercial-Level Video Generation Model in $200k
  • Open-Sora

World Model#

• [2025.08] Genie 3: A new frontier for world models