ecg: https://www.kaggle.com/datasets/devavratatripathy/ecg-dataset ecg5days: https://germain-forestier.info/src/bigdata2018/ (downloaded pre trained models without transfer learning) breast_cancer: https://www.kaggle.com/datasets/yasserh/breast-cancer-dataset stroke: https://www.kaggle.com/datasets/fedesoriano/stroke-prediction-dataset

Tabular data functionality is implemented for "train a new model" and "load pre trained models" case. To be developed is the "import your own model" case.

A basic idea of how whole software works is the pipeline.json files. In there one can observe all the essential information for the dataset, the preprocessing and the training of a model alongside information regarding the counterfactual explanations. When training a new model or using a pre trained one, a pipeline.json file is being used in some way.

On the former case (training a new model) all the information regarding the preprocessing techniques, the test ratio, the name of the dataset and the preprocessed files generated are saved with the goal of being reused when one should load the trained dataset. Therefore, on the latter case, loading a pre trained dataset, this pipeline file is accessed and used to display information for the pre trained model.

For now the following cases are implemented:

1) Train the existed datasets (stroke prediction, breast cancer)
2) Use the pre trained datasets

The layout of the project changed and now everything can be done in one load. For that, ajax requests, javascript and jquery were used to make the whole thing synchronous and replaced the old frontend that would make use of Django template and frontend.