Sometimes LightGBM crashes the jupyter notebook kernel when running in parallel. This needs a bit of investigation, but just using fewer cores seems to resolve that issue for now, so maybe that should be added to the FAQ / docs.
Hi Martin, we ran into trouble when using auto emulate on an apple M1 Mac, working on a conda environment. We fixed it by installing LightGBM via conda instead of pip (conda install lightgbm)
Sometimes LightGBM crashes the jupyter notebook kernel when running in parallel. This needs a bit of investigation, but just using fewer cores seems to resolve that issue for now, so maybe that should be added to the FAQ / docs.