UW-FMRL2/MMMG
收藏Hugging Face2025-05-27 更新2025-10-18 收录
下载链接:
https://hf-mirror.com/datasets/UW-FMRL2/MMMG
下载链接
链接失效反馈官方服务:
资源简介:
MMMG是一个全面且与人类对齐的多模态生成基准,涵盖图像、音频、交替文本和图像、交替文本和音频等4种模态组合。该数据集关注于对生成模型具有重大挑战性的任务,同时能够进行可靠的自动评估。
MMMG is a comprehensive and human-aligned benchmark for multimodal generation across image, audio, interleaved text and image, interleaved text and audio modality combinations. It focuses on challenging tasks for generation models while enabling reliable automatic evaluation.
提供机构:
UW-FMRL2



