SIGSEGV errors and python crashing

#1

At first I received SIGSEGV errors whenever I attempted to create a DMatrix, but downgrading to version 1.2 fixed it. However, now I get SIGSEGVs errors when attempting to train the dataset. No warnings, no nothing, just SIGSEGV error, and python crashing. I have tested on version 1.2 and 1.5, but it didn’t help.

Edit: Training has worked about 5% of the time, as I have attempted to run the code several times.

Edit2: SIGSEGV errors also occurring when attempting to predict.

#2

XGBoost is pretty much unusable on my computer. I have a MacBook Pro from 2016.

#3

There is a problem with the latest libomp package: https://github.com/dmlc/xgboost/issues/7039. It is advised that you downgrade to the previous version of libcomp. See the link for details.