Instructions to use IAMIbrahim/luthorsql-bird-single with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use IAMIbrahim/luthorsql-bird-single with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
LuthorSQL (single-adapter variant)
LoRA adapters and heads on seeklhy/OmniSQL-7B for LuthorSQL, a one-pass text-to-SQL model that never writes SQL token by token.
single/: one adapter that both generates plans (slot heads) and scores executed candidates (rerank head).
Status: private pre-release. The code is in the GitHub repository IAMIbrahimmemon/luthorsql (private).
| EX | |
|---|---|
| DECISION (796 held-out BIRD-train questions) | 58.54 (two-adapter: 59.05) |
| BIRD dev (original gold, 1,534) | 59.84 (two-adapter: 59.39) |
| BIRD hidden test (run by the BIRD team, 1,789) | 68.19 (two-adapter: 68.03) |
On the hidden test: Soft F1 69.35, R-VES 64.58; simple / moderate / challenging EX 77.03 / 65.77 / 43.51.
Speed: 0.84 s per question (median, one H100 PCIe).
Training data: BIRD train, Spider train, a SynSQL-2.5M sample; the validator also used BIRD-Platinum labels, for which no license was found. License: not decided; expected research use only. Limitations: English, foreign-key joins only. The plan grammar covers 84.6% of BIRD dev gold queries.
Citation
Ibrahim Memon, Armin Karamzade, Hanif Jhaveri and Roy Fox. 2026. LuthorSQL: A One-Pass Text-to-SQL Model as System 1 for a Slower LLM Agent. Preprint. Code and data: https://github.com/IAMIbrahimmemon/luthorsql
- Downloads last month
- -
Model tree for IAMIbrahim/luthorsql-bird-single
Base model
seeklhy/OmniSQL-7B