Michal Kurka

FIY: we currently do not use log4j 1.x in any of out standalone builds, we only use it for Hadoop builds and SW builds - both Hadoop and Spark come...

@occhioni-esteco no updates yet - we are currently not investigating this option

@jbentleyEG thank you, is it possible you didn't push `XGLeafAssignMojo`?

@valenad1 this would mean that the issue really only affects small datasets and in reasonably large datasets the values are accurate for all practical purposes, correct?

or is this just about making the test fail and no fix yet?

@jiapei100 thank you for the contribution! ~~@koniecsveta @tomasfryda can you please review?~~

@sebhrusen this cannot be merged until new images are built, correct?

ok, got it - would it be possible to apply the same refactoring/restructuring to the existing code base without changing the versions and then update the versions in a new...

@satai should this be closed?

Shapley is expensive to calculate, can you provide specifics about your model parameters, dataset, and the runtimes?