xgboost交叉验证以及n_jobs报错

今天用xgboost跑了一下分类发生如下报错：

A worker process managed by the executor was unexpectedly terminated.

This could be caused by a segmentation fault while calling the function or by an excessive memory usage causing the Operating System to kill the worker.

 

这是显存不足造成的。

 

解决方案

制定富裕空间的显卡

import os

os.environ["CUDA_VISIBLE_DEVICES"] = "0"

注意，用jupyter发生报错时候可以尝试本地脚本跑！