DeepVision-103K: 一种视觉多样、覆盖面广且可验证的数学数据集,用于多模态推理
DeepVision-103K: A Visually Diverse, Broad-Coverage, and Verifiable Mathematical Dataset for Multimodal Reasoning
AI总结:
DeepVision-103K是一个用于多模态推理的视觉多样、覆盖面广且可验证的数学数据集,通过增强模型的视觉反思和推理能力推动多模态推理发展。
摘要:
基于可验证奖励的强化学习(RLVR)已被证明在增强大多模态模型(LMMs)的视觉反思和推理能力方面是有效的。然而,现有数据集大多源自小规模手动构建或先前资源的重组,这限制了数据的多样性和覆盖面,从而限制了模型性能的进一步提升。为此,我们引入了DeepVision-103K,这是一个用于RLVR训练的综合性数据集,涵盖了多样化的K12数学主题、广泛的知识点和丰富的视觉元素。在DeepVision上训练的模型在多模态数学基准测试中表现强劲,并能有效泛化到一般的多模态推理任务。进一步分析显示,在训练模型中增强了视觉感知、反思和推理能力,验证了DeepVision在推进多模态推理方面的有效性。数据:https://huggingface.co/datasets/skylenage/DeepVision-103K
英文摘要:
Reinforcement Learning with Verifiable Rewards (RLVR) has been shown effective in enhancing the visual reflection and reasoning capabilities of Large Multimodal Models (LMMs). However, existing datasets are predominantly derived from either small-scale manual construction or recombination of prior resources, which limits data diversity and coverage, thereby constraining further gains in model performance. To this end, we introduce \textbf{DeepVision-103K}, a comprehensive dataset for RLVR training that covers diverse K12 mathematical topics, extensive knowledge points, and rich visual elements. Models trained on DeepVision achieve strong performance on multimodal mathematical benchmarks, and generalize effectively to general multimodal reasoning tasks. Further analysis reveals enhanced visual perception, reflection and reasoning capabilities in trained models, validating DeepVision's effectiveness for advancing multimodal reasoning. Data: \href{https://huggingface.co/datasets/skylenage/DeepVision-103K}{this url}.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。