---

• Deploy models in different environments

Input:

• TensorFlow
• scikit-learn
• R
• Spark
• ML Frameworks

Standardized MLFlow model format

Output:

• docker
• Spark
• Serving tools

Directory called mlmodel.

mlflow.model_flavor.save_model(...)  or log_model(...)
mlflow.model_flavor.load_model(...)

• tutorial example