Guiding Distribution Matching Distillation with Gradient-Based Reinforcement Learning
cs.LG, cs.CV
Submitted: 2026-04-21
Updated: 2026-11-02
Project page: https://gdmd-guide.github.io
Terminology
Sources
- Training Diffusion Models with Reinforcement Learning
- Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer
- ShareGPT-4o-Image: Aligning Multimodal Models with GPT-4o-Level Image Generation
- Mean Flows for One-step Generative Modeling
- TempFlow-GRPO: When Timing Matters for GRPO in Flow Models
- Distribution Matching Distillation Meets Reinforcement Learning
- Auto-Encoding Variational Bayes
- Aligning Text-to-Image Models using Human Feedback
- MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE
- BranchGRPO: Stable and Efficient GRPO with Structured Branching in Diffusion Models
- Uniworld-V2: Reinforce Image Editing with Diffusion Negative-aware Finetuning and MLLM Implicit Feedback
- SDXL-Lightning: Progressive Adversarial Diffusion Distillation
- Flow Matching for Generative Modeling
- Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shield
- Flow-GRPO: Training Flow Matching Models via Online RL
- Simplifying, Stabilizing and Scaling Continuous-Time Consistency Models
- Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference
- SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis
- DreamFusion: Text-to-3D using 2D Diffusion
- Progressive Distillation for Fast Sampling of Diffusion Models
Related papers
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis
- (Mis)generalization of Helpful-Only Fine-tuning