program stop training for a middle size graph

Open wenjf opened this issue 6 years ago • 2 comments

Hi, I have applied DynamicGEM on a middle size graph, with 50k nodes and 220k edges. The program can finish the first snapshot training, but then, it appears to be suspended. How can I solve this program?

May 11 '19 13:05 wenjf

Could you provide the specification of your system and the algorithm that was being executed when the program got suspended?

May 13 '19 23:05 Sujit-O

It's a RedHat linux machine, which 128G memory. I just run this program on CPU. After these days' obversion, I found that, sometimes the suspended program will continue to training, sometimes it was just suspended and sleep... Have you ever met this kind of problem. It's really confusing for me...

May 17 '19 03:05 wenjf