# Dataset provenance
- Base: Salesforce/codet5-small (check model license).
- Data: JetBrains-Research/OpenCodeInstruct-Clean (Apache-2.0), CodeXGLUE/CodeSearchNet (C-UDA; underlying repositories may have their own licenses).
- Heuristically selected 15,000 function-derived task examples, not independently verified.
- Inspect original data usage restrictions before redistributing/commercial use.