Issues

Add Constructor to PreviewParseWriter with Domains as Parameter

R: Explaination related functions should check for presence of ggplot2 of a required version

when loading the same data with the same query - it is overwriting the data in memory (changing the frame in memory)

'H2OAutoML' object has no attribute 'explain'

Create better learning curve function in R and Python

Import_sql_select, import_sql_table: fetch mode distributed has data discrepancy then single mode or import_file

GLM Generic model doesn't provide varimp function

Add information about how early stopping is used to tune models in the User Guide

Target Encoding: support feature interaction

Feature Request:uplift trees

create similar feature on Native XGBoost: feature_interaction_constraint

Expose input -> output column mapping in Target Encoder

Kubernetes operator for H2O-3

Fill in Variable Importance user guide page with info about non-tree-based models

Anomaly of H2O AutoML grid-search

POC Add multi-GPU support on single node for XGBoost

Stacked Ensembles don't respect the schema of the base models

XGBoost: Allow CV to use multiple GPUs (on a single node)

Unable to run multiple H2O AutoML processes simultaneously

Connection reset by peer error when running model with > 100000 data rows

Dynamic Stacked Ensemble metalearning strategy in AutoML

StackedEnsemble MOJO to score multiple base models

Deprecate metalearner_* parameters in Stacked Ensemble

Stacked Ensemble: enable using previously persisted base models