CG-Bench

What is CG-Bench?

CG-Bench is meant for evaluating MLLMs' long video understanding, including 12,129 QA pairs from 1219 videos in 3 major question types: perception, reasoning…

Released

2024-12-16

Modality

multimodal

Evaluates

多模态, Multimodal, 理解, Understanding, 多模态模型, VLM, 视频理解, Video Understanding, 不支持, Unsupported

Openness

restricted

Importer

OpenCompass Hub

Reported scores

0

Source provenance

• Original evidence https://hub.opencompass.org.cn/dataset-detail/CG-Bench

CG-Bench paper, code and dataset

• Paper https://arxiv.org/abs/2412.12075v1
• Code https://github.com/CG-Bench/CG-Bench
• Dataset https://huggingface.co/datasets/CG-Bench/CG-Bench

Which sources cite CG-Bench?

• CG-Bench

Related 多模态, Multimodal, 理解, Understanding, 多模态模型, VLM, 视频理解, Video Understanding, 不支持, Unsupported benchmarks

• 3DCodeBench
• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• ActiveFly-Bench
• Ada-LEval
• AECBench