遇见数据集

DAMO-NLP-SG/VideoRefer-Bench

收藏
Hugging Face2024-12-31 更新2025-04-12 收录
官方服务:

资源简介:

VideoRefer Suite是一个用于评估视频对象理解能力的综合基准,包括VideoRefer-Bench-D和VideoRefer-Bench-Q两个子基准。VideoRefer-Bench-D用于评估基于视频的指引用模型在描述生成方面的性能,共有400个数据条目。VideoRefer-Bench-Q用于评估多模态语言模型在解释视频对象方面的熟练程度,包括1000个多项选择题。

VideoRefer Suite is a comprehensive benchmark for evaluating video object understanding capabilities, including two sub-benchmarks VideoRefer-Bench-D and VideoRefer-Bench-Q. VideoRefer-Bench-D is designed to evaluate the description generation performance of video-based referring models with a total of 400 data entries. VideoRefer-Bench-Q is designed to assess the proficiency of multilingual language models in interpreting video objects, including 1000 multiple-choice questions.

提供机构:
DAMO-NLP-SG
二维码
社区交流群
二维码
科研交流群
商业服务