Hi My name is hsuchengmath .
I recently research recommendation system and my Boss told me that
I read MKR the paper . I had read and try to implement it by Pytorch .
But unfortunately , Though I finish code , the accuracy is not acheived it that MKR model should achieve (approximately 50%) .
After frustration , I search for MKR Pytorch on website and find your GitHub .

Can we be friend ?

hi hsientzucheng

i review MRK paper and github code recently for recommendation system
but l can't find negative sampling triples and "scores(positive triple) - scores(negative triple)" loss function
I would be very grateful if you could answer me
thankyou

Hello Author! I'm a newcomer who just paid attention to the recommendation system. I want to ask some questions. I have read your code thoroughly. How can I view the Topk indicators of your code? What is the result of the Topk indicator? For example, recall, what is the result of f1?

Thank the author for his reply and hope to get your reply.

HI，hsientzucheng.When I run your MKR code,It shows that No module named 'trace_grad'，but I can not find trace_grad in Python library. Please help me.thank you!