MAmmoTH-VL/MAmmoTH-VL-Instruct-12M
收藏Hugging Face2025-01-05 更新2024-12-14 收录
下载链接：
https://hf-mirror.com/datasets/MAmmoTH-VL/MAmmoTH-VL-Instruct-12M
下载链接
链接失效反馈官方服务：
资源简介：
MAmmoTH-VL-Instruct-12M数据集是一个视觉指令数据集，其生成流程包括三个步骤：手动数据源收集、使用MLLMs/LLMs重写、以及通过相同的MLLM作为评判者进行过滤。该数据集展示了在数学和科学类别中的详细、逐步响应的示例。
The MAmmoTH-VL-Instruct-12M dataset is generated through a three-step process: manual data source collection, rewriting using Multimodal Large Language Models (MLLMs) or Large Language Models (LLMs), and filtering via the same MLLM. The dataset showcases detailed, step-by-step responses in math and science categories, suitable for large-scale instruction tuning for multimodal reasoning tasks.
提供机构：
MAmmoTH-VL
