Using cupy instead of numpy in xgboost training significantly lowered test-auc score

前端 未结 0 535
爱一瞬间的悲伤
爱一瞬间的悲伤 2020-12-01 11:12

Backgroud:

  • Try to run xgboost model on GPU instead of CPU.
  • Use cupy to replace numpy for better GPU speedup.

Pro

相关标签:
回答
  • 消灭零回复
提交回复
热议问题