[Docs]Update demo readme's use case section with BentoML

[yubozhao]

Hi guys,

I am one of the authors of BentoML. BentoML is a python library for packaging and deploying ML models. It provides high-level APIs for defining an ML service and packaging its artifacts, source code, dependencies, and configurations into a production-system-friendly format that is ready for deployment.

We create this example notebook to showcase the power of XGBoost and how to use it in production setting. We create a simple Titanic survival prediction model with XGBoost, and we are able to use BentoML quickly create a model archive bundle from the Jupyter notebook with all of the dependencies included. Afterward, we start REST API server from generated model archive without writing any web service code.

You can also distribute the model archive as PyPi package, use the built in CLI tool for CI/CD pipeline, create DockerImage with REST server from the generated Dockerfile, or use it as Spark UDF.

Feel free to ping me for any questions on this example and BentoML. We love to hear your feedback and improve the project to help this community.

Cheers

Bo

[hcho3]

Merged. Thanks for the link!