Fast, fully automatic, data-centric out-of-the-box ai model building

Abstract

A system or a method for fast training an artificial intelligence (AI) model. The system accesses a training dataset, divides the training dataset into a training subset and a holdout test subset, and trains the AI model using the training subset. For each of multiple training epochs, the system checkpoints while training the AI model. Checkpointing includes applying the trained AI model to the training subset to determine a train loss value, applying the trained AI model to the holdout test subset to determine a test loss value, and determining whether both train loss values and test loss values converge at a particular epoch. Responsive to determining that convergence occurred at the particular epoch, the system selects the AI model trained at the particular epoch.