UnicEdit-10M：通过统一验证打破规模-质量壁垒的推理增强编辑数据集与基准
UnicEdit-10M: A Dataset and Benchmark Breaking the Scale-Quality Barrier via Unified Verification for Reasoning-Enriched Edits
该论文构建了千万级图像编辑数据集和涵盖空间与知识推理的基准，并提出了非编辑一致性和推理准确率等新指标，为训练和评估图像编辑模型提供了可复用的数据管道与诊断工具。
原文依据 we introduce a lightweight data pipeline that replaces multi-toolchains with an end-to-end model and a unified post-verification stage.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论随着GPT-4o、Nano Banana和Seedream 4.0等强大模型在图像编辑领域的快速发展，闭源与开源模型之间的性能差距正在扩大，主要原因是缺乏大规模、高质量的训练数据以及能够诊断模型在不同编辑行为上弱点的全面基准。现有的数据构建方法面临规模与质量的权衡：人工标注质量高但不可扩展，而自动化流程则存在错误传播和噪声问题。为解决这一问题，我们引入了一种轻量级数据管道，用端到端模型和统一的后验证阶段替代多工具链。为了实现可扩展的质量控制，我们训练了一个70亿参数的双任务专家模型Qwen-Verify，用于高效检测失败和重新描述指令。该管道生成了UnicEdit-10M，一个覆盖多种基础与复杂编辑任务的千万级数据集。我们还提出了UnicBench，一个通用基准，它超越了基础编辑，明确评估空间和知识驱动的推理能力。为了实现细粒度诊断，我们引入了新的指标，包括非编辑一致性和推理准确率。我们在UnicBench上对主流模型的分析揭示了它们的局限性，并为未来研究提供了明确方向。数据集、基准和代码将公开发布。
英文原文摘要 1390 字符
With the rapid advances of powerful multimodal models such as GPT-4o, Nano Banana, and Seedream 4.0 in Image Editing, the performance gap between closed-source and open-source models is widening, primarily due to the scarcity of large-scale, high-quality training data and comprehensive benchmarks capable of diagnosing model weaknesses across diverse editing behaviors. Existing data construction methods face a scale-quality trade-off: human annotations are high-quality but not scalable, while automated pipelines suffer from error propagation and noise. To address this, we introduce a lightweight data pipeline that replaces multi-toolchains with an end-to-end model and a unified post-verification stage. For scalable quality control, we train a 7B dual-task expert model, Qwen-Verify, for efficient failure detection and instruction recaptioning. This pipeline yields UnicEdit-10M, a 10M-scale dataset spanning diverse basic and complex editing tasks. We also propose UnicBench, a general benchmark that extends beyond basic edits to explicitly assess spatial and knowledge-driven reasoning. To enable fine-grained diagnosis, we introduce novel metrics, including Non-edit Consistency and Reasoning Accuracy. Our analysis of mainstream models on UnicBench reveals their limitations and provides clear directions for future research. The dataset, benchmark, and code will be released.