DICS：探索数据内在一致性用于视觉指令选择

摘要

视觉指令微调对于提升视觉语言模型的视觉-语言对齐和指令跟随能力至关重要。然而，在固定比例约束下从快速扩展的数据集中识别最优子集仍然是一个重大瓶颈。现有方法主要依赖分布多样性或启发式过滤，但往往忽视了单个样本内部的连贯性。为弥补这一不足，我们提出数据内在一致性（DIC），一种用于量化样本级组件间一致性的自评分指标。DIC包含两个模块：视觉信息一致性（VIC），评估视觉内容与指令之间的对齐程度；以及响应信息一致性（RIC），评估响应相对于指令的连贯性。基于DIC，我们进一步提出数据内在一致性选择（DICS），一种自适应数据选择方法，在不同数据预算下优化高样本内一致性与全局分布多样性之间的权衡。大量实验表明，DICS在多种数据集规模和模型架构上持续优于最先进方法，仅使用LLaVA-1.5-665K数据的25%即可超越全量数据集微调的性能。我们进一步构建了DICS-6M，一个包含600万样本的多模态指令语料库，推动了迄今为止最大规模的视觉指令选择研究；值得注意的是，DICS仅使用其报告训练数据的不到25%即可达到官方InternVL3-8B-Instruct性能的94.52%。代码可在https://github.com/cqu-student/DICS获取。

English

Visual instruction tuning is crucial for advancing the vision-language alignment and instruction-following capabilities of Vision-Language Models (VLMs). However, identifying optimal subsets under a fixed ratio constraint from rapidly expanding datasets remains a significant bottleneck. While existing methods largely depend on distribution diversity or heuristic filtering, they often overlook the internal coherence within individual samples. To bridge this gap, we propose Data Intrinsic Consistency (DIC), a self-scoring metric designed to quantify the sample-level inter-component consistency. DIC consists of two modules: Visual Information Consistency (VIC), evaluating the alignment between visual content and instructions, and Response Information Consistency (RIC), assessing response coherence relative to the instruction. Building upon DIC, we introduce Data Intrinsic Consistency Selection (DICS), an adaptive data selection method that optimizes the trade-off between high intra-sample consistency and global distributional diversity under varying data budgets. Extensive experiments demonstrate that DICS consistently outperforms state-of-the-art methods across diverse dataset scales and model architectures, surpassing full-dataset fine-tuning while using only 25% of the LLaVA-1.5-665K data. We further curate DICS-6M, a 6M-sample multi-modal instruction corpus that enables the largest-scale visual instruction selection study to date; remarkably, DICS reaches 94.52\% of the official InternVL3-8B-Instruct performance using less than 25\% of its reported training data. Code can be seen at https://github.com/cqu-student/DICS