SEMODS: A Validated Dataset of Open-Source Software Engineering Models (March 2026 Snapshot)

This dataset accompanies the paper "SEMODS: A Validated Dataset of Open-Source Software Engineering Models", accepted at the 3rd ACM International Conference on AI Foundation Models and Software Engineering (FORGE 2026).

Dataset Description 

SEMODS (Software Engineering MODels) is a validated dataset comprising 4,132 open-source software engineering models hosted on Hugging Face. It provides a structured view of the SE model ecosystem as of March 2026, addressing the lack of SE-specific catalogues in generalist model registries. It offers a standardized representation of model evaluation results and supports multiple applications such as data analysis, model discovery, benchmarking, and model adaptation. 

The dataset is provided as a PostgreSQL dump file. Important: The dump was created using PostgreSQL 17. Users must restore it using PostgreSQL server and client tools version 17 or newer. Dumps created with newer PostgreSQL versions are not backward compatible.

How to load the database locally

Install and start PostgreSQL 17:

brew install postgresql@17

brew services start postgresql@17

     2. Create the database:

/opt/homebrew/opt/postgresql@17/bin/createdb semods

     3. Restore the dump (Replace /path/to/ with the actual path to your downloaded file)

/opt/homebrew/opt/postgresql@17/bin/pg_restore -d semods /path/to/SEMODS_202511.dump

Alternatively, you can use graphical tools such as DBeaver by connecting to a PostgreSQL 17 instance and using the "Restore" option.

Citation

Please cite the associated FORGE 2026 paper when using this dataset:

@article{gonzalez2026semods,  title={SEMODS: A Validated Dataset of Open-Source Software Engineering Models},  author={Gonz{\'a}lez, Alexandra and Franch, Xavier and Mart{\'\i}nez-Fern{\'a}ndez, Silverio},  journal={arXiv preprint arXiv:2601.00635},  year={2026}}