Shentong Mo
2 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
The paper proposes VRPO, a reinforcement learning-based optimization strategy that replaces static alignment losses in diffusion models, significantly improving both convergence and image fidelity.
The paper proposes VRPO, a reinforcement learning-based optimization strategy that replaces static alignment losses in diffusion models, significantly improving both convergence and image fidelity.
Papers
Improving Visual Representation Alignment Generation with GRPO
The paper proposes VRPO, a reinforcement learning-based optimization strategy that replaces static alignment losses in diffusion models, significantly improving both convergence and image fidelity.