An ML research team creates custom ML models and shares the model artifacts with other teams for integration. The ML team keeps the training code and data. They need to establish a mechanism for the ML team to audit the models.

What solution should the ML team implement when publishing the custom ML models?