AI Model Collection

Barroso, M., & Daniel, H. G. (2024). AI Model Collection (1.0.0). Zenodo. https://doi.org/10.5281/zenodo.10957139

Python AI Model classes used in the context of Knowledge EU Project

The models implemented are:

• Autogluon tabular model
• Classifier and Regressor sklearn models, lightgbm, xgboost, catboost
• Tensorflow keras deep learning models using different architectures, pretrained NN and finetuning strategies.

All the models inherit from a general abstract class that implements the following methods:

• fit
• predict
• evaluate
• _load_model
• _save_model
• to_pmml
• from_pmml
• to_onnx
• from_onnx
• get_metadata

All the models can be converted to PMML or ONNX standards.