Datasets:
Multi-MME
Multi-MME is the multilingual version of MME, a comprehensive yes/no benchmark for evaluating multimodal large language models. It is released in English, Italian, French, German and Spanish, and contains 11,870 entries (2,374 per language).
The dataset is part of MEVBench (Multimodal European Vision Benchmark) and was used as an evaluation benchmark in the paper Multilingual Training and Evaluation Resources for Vision-Language Models.
Source dataset
| Dataset | MME |
| Hugging Face | darkyarding/MME |
| Paper | MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models (Fu et al., NeurIPS 2023 Datasets and Benchmarks Track) |
MME is a comprehensive yes/no benchmark measuring both perception (10 subtasks) and cognition (4 subtasks), with minimal instructions followed by "Please answer yes or no." so that all models are compared under identical conditions. The Hugging Face release used here (darkyarding/MME) contains 2,374 instruction–answer pairs over 1,187 images, balanced between 1,187 "Yes" and 1,187 "No".
How the dataset was built
The English benchmark was extended to four additional languages through an automated translation pipeline, keeping the evaluation structure intact:
- Translated fields:
question. - Translation model:
gpt-4.1-mini. - Unchanged: images,
question_id, the goldanswerand thecategory(subtask) label, so that per-subtask scores remain comparable with the English benchmark. - English configuration: the
enconfiguration contains the original English questions and is kept as a reference for cross-lingual comparison. - No regeneration or re-annotation: only the question text was translated; images were not modified or regenerated.
The translation protocol is described in Section 4 of the accompanying paper.
Dataset structure
Configurations and splits
Each language is a separate configuration, and all configurations share the same split name (test).
| Config | Language | Split | Entries |
|---|---|---|---|
en |
English | test | 2,374 |
fr |
French | test | 2,374 |
de |
German | test | 2,374 |
it |
Italian | test | 2,374 |
es |
Spanish | test | 2,374 |
| Total | 11,870 |
Fields
| Field | Description |
|---|---|
question_id |
Identifier inherited from the source release |
image |
Image associated with the question |
question |
Question text (translated) |
answer |
Gold answer: Yes or No |
category |
MME subtask, e.g. existence, count, OCR, commonsense_reasoning (unchanged, in English) |
Usage
from datasets import load_dataset
ds = load_dataset("VillanovaAI/Multi-MME", "it", split="test")
print(ds[0]["question"])
print(ds[0]["answer"], ds[0]["category"])
Limitations
- The translations are produced automatically and inherit any error of the source benchmark as well as any error of the translation model. They are intended for evaluation, not for training.
Licensing and attribution
This multilingual release is distributed under Apache 2.0, coherently with the license declared for the Hugging Face release used as source, and the upstream benchmark is MME (Fu et al., NeurIPS 2023 Datasets and Benchmarks Track). Images are redistributed unchanged; only the question text is translated.
Citation
If you use this dataset, please cite the paper describing the multilingual benchmark suite and the original MME paper.
@article{baiamonte2026multilingual,
title={Multilingual Training and Evaluation Resources for Vision-Language Models},
author={Baiamonte, Daniela and Fano, Elena and Gabburo, Matteo and Simonazzi, Stefano and Rigutini, Leonardo and Zugarini, Andrea},
journal={arXiv preprint arXiv:2604.18347},
year={2026}
}
@inproceedings{fu2023mme,
title={MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models},
author={Fu, Chaoyou and Chen, Peixian and Shen, Yunhang and Qin, Yulei and Zhang, Mengdan and Lin, Xu and Yang, Jinrui and Zheng, Xiawu and Li, Ke and Sun, Xing and Wu, Yunsheng and Ji, Rongrong and Shan, Caifeng and He, Ran},
booktitle={Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track},
year={2023}
}
- Downloads last month
- 119