aws-samples/amazon-sagemaker-custom-loss-function

This code shows how to train a model in Amazon SageMaker using a custom loss function for a binary classification problem in which the costs of different kinds of misclassification are very different.

Jupyter NotebookMIT-0