The MLeap Runtime is a lightweight execution engine for machine learning pipelines. It has the following features:
- Data frames, called leap frames with support for all common data types and custom data types
- Transformers, currently supports all transformers found in Spark as well as multiple extension transformers
- Pipelines, easily build pipelines from your transformers
- Full integration with MLeap Bundles, the MLeap runtime provides a reference implementation for MLeap Bundles and anyone who wishes to implement their own serializers for it
- Serialization formats for leap frames to easily send content across the wire
- A very fast linear algebra system powered by BLAS
See our MLeap Runtime usage section for more information on how to use the MLeap Runtime in your application.