MME-RealWorld Benchmark

What is MME-RealWorld?

MME-RealWorld evaluates MLLMs' real-world recognition, featuring 13,366 high-resolution images averaging 2,000 × 1,500 pixels.

Released

2024-08-23

Modality

multimodal

Evaluates

多模态, Multimodal, 理解, Understanding, 多模态模型, VLM, 图像理解, Image Understanding, 不支持, Unsupported

Openness

unknown

Importer

OpenCompass Hub

Reported scores

0

Source provenance

• Original evidence https://hub.opencompass.org.cn/dataset-detail/MME-RealWorld

MME-RealWorld paper, code and dataset

• Paper https://arxiv.org/abs/2408.13257
• Code https://github.com/yfzhang114/MME-RealWorld
• Dataset https://huggingface.co/datasets/yifanzhang114/MME-RealWorld
• Dataset https://huggingface.co/datasets/yifanzhang114/MME-RealWorld-Base64

Which sources cite MME-RealWorld?

• MME-RealWorld

Related 多模态, Multimodal, 理解, Understanding, 多模态模型, VLM, 图像理解, Image Understanding, 不支持, Unsupported benchmarks

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