從語料庫到共同演化能力：以能力為中心的通用圖像生成資料設計

摘要

大規模影像生成已受益於資料規模、品質、重新平衡與重新生成描述等方面的進展，然而傳統流程通常孤立地最佳化特定任務的資料集。核心挑戰不僅在於如何整理每個任務特定的語料庫，還在於如何根據生成能力之間的依賴關係來組織異質監督。我們提出一個以能力為驅動的資料基礎設施，將特定能力的監督建構與對齊能力的課程排程相結合。其中的三個專業但可互操作的資料引擎，分別為文字-影像對應、影像間轉換與影像-知識關聯建構互補的關係監督；同時，描述專家使文字到影像（T2I）與編輯監督在不同任務與粒度上保持一致。一個多階段課程依循能力習得的依賴順序，共同演進任務組成、視覺概念分布、資料品質與影像解析度；能力感知評估則透過針對性檢索、專家建構與差距感知的重新取樣來閉環。在大規模應用中，該框架整理了包含 4.4 億張影像的 T2I 語料庫、1.2 億個編輯對，以及超過 2700 萬個影像-實體對。利用此基礎設施，我們從零開始訓練兩種規模的多模態擴散模型，參數量分別為 3B 與 6B。我們在 CPI-Bench 上進行定量評估，並在各種文字到影像與編輯場景中進行定性評估。實驗結果展現了廣泛的視覺覆蓋、多功能的渲染，以及跨生成能力的有效遷移。

English

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 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.