search

zhengruifeng / spark-libFM

An implement of Factorization Machines (LibFM)
Apache License 2.0
248 stars 119 forks source link

How to deal with the NAN in the weightVector #16

Open zzwzhao opened 6 years ago

EasonYan commented 5 years ago

in my case, NAN value happens since com.intel.imllib.fm.regression.FMGradient.predict generates Infinity value, get stuck too