search

blebreton / spark-FM-parallelSGD

Implementation of Factorization Machines on Spark using parallel stochastic gradient descent (python and scala)
Apache License 2.0
229 stars 80 forks source link

error in fm_get_p function #8

Open wsgtcyx opened 5 years ago

wsgtcyx commented 5 years ago

Hi, I found that in the 51th line of your fm_parallel_sgd.scala file, the result of "val phi = 0.5(VX:VX - VX_square).sum" and "val phi = 0.5(VX:VX :- VX_square).sum" is different, which the second one should be right, described in original paper. please check. Thank you!

luweiagi commented 4 years ago

Hi, I found that in the 51th line of your fm_parallel_sgd.scala file, the result of "val phi = 0.5(VX:_VX - VXsquare).sum" and "val phi = 0.5(VX:VX :- VX_square).sum" is different, which the second one should be right, described in original paper. please check. Thank you!

I think the author is right.