In order to run at NERSC I need to change n_jobs=-1 to n_jobs=1 in src/flexcode/regression_models.py

If I don't make this change, in crashes with 

`xgboost.core.XGBoostError: [10:41:13] /workspace/src/data/array_interface.cu:44: Check failed: err == cudaGetLastError() (0 vs. 2)`


It would be great if either, 

We had some guidance on how to avoid this crash in general, 
or 
we could make the n_jobs a parameter that gets picked up somewhere, to make it easier to switch when running at NERSC