MMR1/MMR1-RL
收藏Hugging Face2025-10-01 更新2025-10-18 收录
下载链接：
https://hf-mirror.com/datasets/MMR1/MMR1-RL
下载链接
链接失效反馈官方服务：
资源简介：
MMR1数据集是一个包含约160万长链式思维（CoT）冷启动轨迹和约1.5万强化学习QA对的大型多模态推理数据集。这些数据覆盖了数学、科学、图表/图形、文档表格和一般理解等多个领域，结合了现有的公共资源（如MathVerse、ScienceQA、ChartQA、DocVQA、GQA）以及新策划和自收集的数据，确保了数据的质量、难度和多样性。
The MMR1 dataset is a large-scale multimodal reasoning dataset containing about 1.6 million long Chain-of-Thought (CoT) cold-start trajectories and about 15,000 reinforcement learning QA pairs. These data cover multiple domains including mathematics, science, charts/figures, document tables, and general understanding, integrating existing public resources (such as MathVerse, ScienceQA, ChartQA, DocVQA, GQA) with newly curated and self-collected data to ensure quality, difficulty, and diversity.
提供机构：
MMR1
