From Corpora to Co-Evolving Capabilities: Capability-Centric Data Design for Generalist Image Generation
초록
Large-scale image generation has benefited from advances in data scale, quality, rebalancing, and recaptioning, yet conventional pipelines typically optimize task-specific datasets in isolation. A central challenge is not only how to curate each task-specific corpus, but also how to organize heterogeneous supervision according to the dependencies among generative capabilities. We present a \textbf{capability-driven data infrastructure} that couples capability-specific supervision construction with capability-aligned curriculum scheduling. Its three specialized yet interoperable data engines build complementary relational supervision for text-image grounding, inter-image transformation, and image-knowledge association, while caption experts align T2I and editing supervision across tasks and granularities. A multi-stage curriculum jointly evolves task composition, visual-concept distribution, data quality, and image resolution along the dependency order of capability acquisition, with capability-aware evaluation closing the loop through targeted retrieval, expert construction, and gap-aware resampling. At scale, the framework curates a 440M-image T2I corpus, 120M editing pairs, and over 27M image-entity pairs. With this infrastructure, we train multimodal diffusion models at two scales from scratch, with 3B and 6B sizes respectively. We conduct quantitative evaluation on CPI-Bench, along with qualitative evaluations across diverse text-to-image and editing scenarios. Experimental results present broad visual coverage, versatile rendering, and effective transfer across generative capabilities.
저자 (17명)
- Xingjian Wang — LinkedIn 검색
- Zhao Wang — LinkedIn 검색
- Taihang Hu — LinkedIn 검색
- Jun Zheng — LinkedIn 검색
- Qing Jin — LinkedIn 검색
- Qinye Zhou — LinkedIn 검색
- Zhengtao Wu — LinkedIn 검색
- Yongchao Du — LinkedIn 검색
- Zuan Gao — LinkedIn 검색
- Chao Lin — LinkedIn 검색
- Yefeng Shen — LinkedIn 검색
- Xiaoli Xu — LinkedIn 검색
- Zhengze Xu — LinkedIn 검색
- Hao Yan — LinkedIn 검색
- Yuhang Yu — LinkedIn 검색
- Mingzhou Zhang — LinkedIn 검색
- Mengting Chen — LinkedIn 검색
저자 LinkedIn 변경 추적은 추후 자동화 예정입니다. 현재는 검색 링크를 제공합니다.