Sample By Step, Optimize By Chunk: Chunk-Level GRPO For Text-to-Image Generation
This is an incremental improvement for text-to-image generation methods, addressing specific limitations in GRPO-based approaches.
The paper tackled inaccurate advantage attribution and neglect of temporal dynamics in Group Relative Policy Optimization (GRPO) for text-to-image generation by shifting optimization from step to chunk level, resulting in superior preference alignment and image quality.
Group Relative Policy Optimization (GRPO) has shown strong potential for flow-matching-based text-to-image (T2I) generation, but it faces two key limitations: inaccurate advantage attribution, and the neglect of temporal dynamics of generation. In this work, we argue that shifting the optimization paradigm from the step level to the chunk level can effectively alleviate these issues. Building on this idea, we propose Chunk-GRPO, the first chunk-level GRPO-based approach for T2I generation. The insight is to group consecutive steps into coherent 'chunk's that capture the intrinsic temporal dynamics of flow matching, and to optimize policies at the chunk level. In addition, we introduce an optional weighted sampling strategy to further enhance performance. Extensive experiments show that ChunkGRPO achieves superior results in both preference alignment and image quality, highlighting the promise of chunk-level optimization for GRPO-based methods.