An End-to-End Ukrainian RAG for Local Deployment. Optimized Hybrid Search and Lightweight Generation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915953066901504 |
|---|---|
| author | Trokhymovych, Mykola Oliinyk, Yana Nyzhnyk, Nazarii |
| author_facet | Trokhymovych, Mykola Oliinyk, Yana Nyzhnyk, Nazarii |
| contents | This paper presents a highly efficient Retrieval-Augmented Generation (RAG) system built specifically for Ukrainian document question answering, which achieved 2nd place in the UNLP 2026 Shared Task. Our solution features a custom two-stage search pipeline that retrieves relevant document pages, paired with a specialized Ukrainian language model fine-tuned on synthetic data to generate accurate, grounded answers. Finally, we compress the model for lightweight deployment. Evaluated under strict computational limits, our architecture demonstrates that high-quality, verifiable AI question answering can be achieved locally on resource-constrained hardware without sacrificing accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_22095 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | An End-to-End Ukrainian RAG for Local Deployment. Optimized Hybrid Search and Lightweight Generation Trokhymovych, Mykola Oliinyk, Yana Nyzhnyk, Nazarii Computation and Language This paper presents a highly efficient Retrieval-Augmented Generation (RAG) system built specifically for Ukrainian document question answering, which achieved 2nd place in the UNLP 2026 Shared Task. Our solution features a custom two-stage search pipeline that retrieves relevant document pages, paired with a specialized Ukrainian language model fine-tuned on synthetic data to generate accurate, grounded answers. Finally, we compress the model for lightweight deployment. Evaluated under strict computational limits, our architecture demonstrates that high-quality, verifiable AI question answering can be achieved locally on resource-constrained hardware without sacrificing accuracy. |
| title | An End-to-End Ukrainian RAG for Local Deployment. Optimized Hybrid Search and Lightweight Generation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2604.22095 |