The models described here were developed in jupyter-lab, and the drug dataset were compiled from various authenticated resources such as DrugBank, DrugCentral, federal register, etc., A total of 2,212 compounds consisting of approved and withdrawn drugs were cleaned, processed and features such as the molecular properties and fingerprints were trained on five scalable machine learning algorithms. These models can be used to predict the safety or toxicity of any molecule, based on the computed molecular features, and can be included in the pipeline of computational drug discovery process. For details, please refer to John, L.; Mahanta, H. J.; Soujanya, Y.; Sastry, G. N. Assessing Machine Learning Approaches for Predicting Failures of Investigational Drug Candidates during Clinical Trials. Computers in Biology and Medicine 2023, 153, 106494. https://doi.org/10.1016/j.compbiomed.2022.106494.