ZejunLi/MoVT-Train

Mixture-of-Visual-Thoughts训练数据集是用于视觉推理的上下文自适应推理模式选择研究的训练数据。数据集包括两种思考模式的数据：基于文本的思考和视觉定位的思考，以及几何、对象计数、OCR、图表、定位、科学等领域的问题数据。Mixture-of-Visual-Thoughts Training Data is the training dataset for research on context-adaptive reasoning mode selection for visual reasoning. The dataset includes data for two thinking modes: text-based thinking and visually-grounded thinking, as well as data for questions in geometry, object counting, OCR, Chart, Grounding, Science...