OpenGVLab/MMPR-v1.1
收藏Hugging Face2025-04-13 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/OpenGVLab/MMPR-v1.1
下载链接
链接失效反馈官方服务：
资源简介：
MMPR是一个包含约300万个样本的大规模、高质量多模态推理偏好数据集。该数据集通过混合偏好优化方法（MPO）用于微调InternVL2-8B模型，显著提升了模型在多模态推理任务上的性能。
MMPR is a large-scale, high-quality multimodal reasoning preference dataset containing about 3 million samples. This dataset is used to fine-tune the InternVL2-8B model with the Mixed Preference Optimization (MPO) method, significantly improving the models performance on multimodal reasoning tasks.
提供机构：
OpenGVLab
