---
license: cc-by-nc-4.0
language:
- ro
---

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->
[HoraVQA](https://huggingface.co/datasets/CohereLabs/AyaVisionBench) is a benchmark designed to evaluate vision-language models understanding of Romanian culture. Images are crawled and curated by humans, while questions and answer are built exclusively by human annotators.

This dataset is used as a benchmark and is part of the evaluation protocol for Romanian VLMs proposed in *"Înțelegi românește?" A Recipe for Romanian Vision-Language Models* ([Masala et al., 2026](https://arxiv.org/abs/2605.31401))


## Citation

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

```bibtext
@misc{masala2026inctelegiromanecstereciperomanian,
      title={"\^{I}n\c{t}elegi Rom\^ane\c{s}te?'' A Recipe for Romanian Vision-Language Models}, 
      author={Mihai Masala and Marius Leordeanu and Mihai Dascalu and Traian Rebedea},
      year={2026},
      eprint={2605.31401},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2605.31401}, 
}
```