Hi authors,
thank you for your great work on OpenGraph and for making the code publicly available!
I am currently trying to reproduce the results from Table 1 in your paper, but I've noticed the number of pre-training epochs isn't specified .
I'm reproducing the paper results for 50 epochs (default), my R@20 results are: ML-1M=0.1934 (paper: 0.1911✓), ML-10M=0.2191 (paper: 0.2370), collab=0.0434 (paper: 0.0421✓).
Could you clarify:
(1) How many epochs were actually used for the paper's results?
(2) How to calculate R@40 - should I just run with --topk 40 flag?
(3) Any suggestions for closing the ML-10M performance gap?
Any guidance would be greatly appreciated.
Thank you for your time and help! 🙏
Hi authors,
thank you for your great work on OpenGraph and for making the code publicly available!
I am currently trying to reproduce the results from Table 1 in your paper, but I've noticed the number of pre-training epochs isn't specified .
I'm reproducing the paper results for 50 epochs (default), my R@20 results are: ML-1M=0.1934 (paper: 0.1911✓), ML-10M=0.2191 (paper: 0.2370), collab=0.0434 (paper: 0.0421✓).
Could you clarify:
(1) How many epochs were actually used for the paper's results?
(2) How to calculate R@40 - should I just run with --topk 40 flag?
(3) Any suggestions for closing the ML-10M performance gap?
Any guidance would be greatly appreciated.
Thank you for your time and help! 🙏