MMR1/MMR1-SFT

MMR1-SFT是一个大规模的、经过精心策划的视觉-语言长链式思维（CoT）数据集，用于冷启动的多模态推理模型的监督微调。该数据集包含约160万个带有经过验证的长CoT推理和简短答案的多模态问答示例。数据集的领域主要包括数学（大多数）、科学、图表/图形、文档/表格和一般领域。数据集的目的是为推理导向的多模态大型语言模型（MLLM）提供一个开放的、可复制的冷启动语料库，具有长的链式思维。数据集的结构包括对话和图像两部分，对话部分包含人类和gpt的交流回合，图像部分包含与视觉内容相关的相对路径。

MMR1-SFT is a large-scale, carefully curated vision-language long chain-of-thought (CoT) dataset for cold-start supervised fine-tuning of multimodal reasoning models. The dataset contains approximately 1.6 million multimodal QA examples with verified long CoT rationales and short answers. The domains include Math (majority), Science, Chart/Figure, Doc/Table, and General. The purpose of the dataset is to provide an open, reproducible cold-start corpus with long CoTs for reasoning-oriented MLLMs. The data structure includes conversations and images, where conversations contain human and gpt dialogue turns, and images include relative paths to associated visual content.