fid

This paper studies the trade-offs in LoRA rank selection for diffusion model fine-tuning, showing that small-to-moderate ranks like 4 and 8 optimize efficiency by balancing FID scores and computational costs.

This paper presents a reinforcement learning framework for visual generative models that uses distribution-wise rewards, with a subset-replace strategy for efficiency, improving image diversity and quality while addressing mode collapse and reward hacking.

This paper analyzes the variance of FID scores across different training and sampling seeds, revealing significant reproducibility issues in image generation evaluation. It proposes a new evaluation protocol with error bars and per-cell optimal guidance tuning.