- Distirbuted trianing TF (1.x and 2.0)
- Distirbuted training for LightGBM
- TF-Serving as an option for delpolyment in addition to Seldon model server
- Improve documentation and testing
- Improved support for Append builder
- Support for distributed training in other frameworks: PyTorch, Spark ML, etc.
- Integrations: KF-Serving, KF-Pipelines, Katib
- Integration tests for on-prem use case