Plot tuner data with Area under Learning Curve (with error bars) + Fix in majority voting

google / nitroml

NitroML is a modular, portable, and scalable model-quality benchmarking framework for Machine Learning and Automated Machine Learning (AutoML) pipelines.

Apache License 2.0

42 stars 6 forks source link

Plot tuner data with Area under Learning Curve (with error bars) + Fix in majority voting #47

Closed nikhil-dce closed 4 years ago

nikhil-dce commented 4 years ago

Plot tuner data with ALC + error bars. (See fig. below)
Fix in majority-voting. Added an additional test that checks the output generated by metalearner.
Added documentation for MLMD connections when working with Kubeflow.

result