Philip Hyunsu Cho

@trivialfis Should we pin hypothesis to `

Also ensure that the local and online machines have the same processor architecture. There could be an issue with endianness.

I don't think XGBoost has been tested with Pyinstaller. If anyone else can jump in and propose a fix, that would be great.

Using multi:softmax objective will produce class predictions. For now, please use multi:softprob. I would welcome your suggestion for improving the behavior. I can think of following options: 1. Document the...

No concrete estimate yet. We are short-handed right now, especially for the JVM and R packages.

Also, Spark 3.0 is built with Scala 2.12. The policy is that XGBoost JVM package uses the version of Scala that's used by Spark. As a result, we probably won't...

@imkunals726 Are you able to share your model file?

For the record, here is the command to launch a K8s cluster on GCP: ``` gcloud container clusters create phcho-rapids-gpu-kubeflow \ --accelerator type=nvidia-tesla-p100,count=2 --machine-type n1-standard-4 \ --zone us-central1-c --release-channel stable...

Support for Scala 2.13 was added in #9099

Fixed in the latest cuML