Metric-AI/ms_marco_hye

---
language:
- hy
task_categories:
- text-retrieval
license: cc-by-nc-4.0
---
# Armenian MS MARCO Dataset
This dataset is part of [ArmBench-TextEmbed (Armenian Embedding Model Benchmark)](https://huggingface.co/spaces/Metric-AI/ArmBench-TextEmbed).
We translated a 10K subset of the [MS MARCO (Bajaj et al., 2018) validation set (v2.1)](https://huggingface.co/datasets/microsoft/ms_marco/viewer/v2.1) using Gemini 2.0 Flash.
This dataset is used for Armenian retrieval evaluation tasks.

# Citation
If you use this dataset in your research, please cite our paper:

```bibtex
@inproceedings{navasardyan2026less,
title={Less is More: Adapting Text Embeddings for Low-Resource Languages with Small Scale Noisy Synthetic Data},
author={Navasardyan, Zaruhi and Bughdaryan, Spartak and Minasyan, Bagratuni and Davtyan, Hrant},
booktitle={Proceedings of the Workshop on Low-Resource Language Models (LoResLM) at EACL 2026},
year={2026},
url={https://arxiv.org/abs/2603.22290}
}
```

## License
This dataset is a translation/derivative of [MS MARCO](https://huggingface.co/datasets/microsoft/ms_marco) and is licensed under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/). Use is restricted to non-commercial research purposes in accordance with the [Microsoft Research License Terms](https://microsoft.github.io/msmarco/)