New to using H2O.ai so please bear with me...
I'm running H2O.ai, working via the Web UI for simplicity sake and mostly experimenting with the AutoML features. My understanding is that H2O4GPU is a separate package.
Is there any integration between the WebUI functionality, specifically using AutoML, with H2O4GPU? Or do I have to manually code via Python first, train, then export the model and import into the WebUI for future use?
The purpose of this question is that I'm debating on whether or not I should build a GPU enabled Linux machine to run these workloads, but if there is no integration between the WebUI and AutoML amd H2O4GPU (without having to manually code) then I'll just keep running my training on CPUs.
Any information would be appreciated. Unfortunately I couldn't find any documentation regarding this type of integration between the two other than it's a separate package. I'm assuming the standard download of the community version of H2O.ai doesn't work with GPUs.
There is no integration currently, however the xgboost algorithm itself is now embedded in h2o-3 directly.