Tool for estimating desired H2O cluster size based on input data

Description

We would like to have a mechanism/tool that would be able to estimate how big should H2O cluster be for a given dataset. This can be a wizard that asks what kind of algos will the user use (automl/gbm,...) and ask for the input dataset/datasets. Based on this input and environment restrictions (max memory per node/container) we will provide a guidance how to configure the cluster.

Assignee

New H2O Bugs

Reporter

Michal Kurka

Support ticket URL