why stop_gradient - Githubissues

hexiangnan / adversarial_personalized_ranking

Adversarial Learning, Matrix Factorization, Recommendation

215 stars 62 forks source link

Open nickhuangxinyu opened 5 years ago

nickhuangxinyu commented 5 years ago

Hi xiangnan.

why you use stop_gradient? just use gradient is not ok?

b3326023 commented 4 years ago

Same question here