A collection of benchmarks for collaborative learning.
AgoraLearn is a research-oriented Python benchmark suite focused on collaborative learning, in which models enhance their predictive performance by leveraging training data from other related datasets to improve generalization.
- Official source code repo: https://github.com/hcherkaoui/agoralearn
- MIT License: https://opensource.org/license/mit
In order install the package, run:
git clone https://github.com/hcherkaoui/agoralearn
cd agoralearn
pip install -r requirements.txt
pip install -e .The required dependencies to use the software are:
- numba (>=0.61.2)
- numpy (>=2.2.6)
- Pandas (>=2.3.1)
- Torch (>=2.7.1)
- Torchvision (>=0.22.1)
- Transformers (>=4.53.1)
- Joblib (>=0.16.0)
- Scikit-learn (>=1.7.0)
- Seaborn (>=0.13.2)
- Matplotlib (>=3.10.0)
In order to launch the unit-tests, run the command:
pytest # run the unit-testsIn order to check the PEP 8 compliance level of the package, run the command::
flake8 --ignore=E501,W503 --count agoralearn