BytedanceDouyinContent/VGR
收藏Hugging Face2026-03-10 更新2025-07-05 收录
下载链接：
https://hf-mirror.com/datasets/BytedanceDouyinContent/VGR
下载链接
链接失效反馈官方服务：
资源简介：
VGR-SFT（视觉定位推理-监督微调）是一个大规模的多模态推理数据集，与论文“VGR: Visual Grounded Reasoning”相关。该数据集是首次尝试在多模态推理中显式建模视觉区域注意力，包含了混合视觉定位和语言推理的推理数据。它使模型能够在推理过程中自主关注任意的视觉区域。数据集包括图像、问题、推理链和答案，以及与推理相关的视觉区域注释。
VGR-SFT (Visual Grounded Reasoning - Supervised Fine-Tuning) is a large-scale multimodal reasoning dataset associated with the paper VGR: Visual Grounded Reasoning. This dataset represents the first attempt to explicitly model visual region attention in multimodal reasoning, containing reasoning data with mixed vision grounding and language deduction. It enables models to autonomously attend to arbitrary visual regions during the reasoning process, including images, questions, reasoning chains, and answers with annotations of visual regions relevant to the reasoning.
提供机构：
BytedanceDouyinContent
