bovane
commented
2 days ago I'm sorry, it was my problem; my experimental environment was insufficient, which caused the training and testing times to be very long, and it wasn't a problem with the code.
However, there is one issue: why is the CPU usage very high during encoding and decoding, while the GPU usage is quite low?
Hello Malik
Recently, I have been very interested in research related to large-scale knowledge graphs, so I found your published work. I am very intrigued by your innovations, so I read your paper and tried to run the code you provided. However, I encountered some issues. The details are as follows:
Experimental Environment:
Ubuntu 22.04 Miniconda -> virtualenv python=3.8 GPU: RTX 4090D (24GB) * 1 CPU: 15 vCPU Intel(R) Xeon(R) Platinum 8474
According to the README file in the code repository, I was able to successfully proceed to the "Preparing Clusters" step. However, during the "Running Encoder" step, the program gets stuck for over ten minutes without any changes. (Detailed information, as shown in the image below.)
When I checked the resource usage, the CPU usage reached 800%, while the GPU usage was only 30%.
In this case, is there an issue with my environment configuration, or is there a problem with my operation? Please help me. I followed all the instructions in the README, and I executed the commands without any changes.
I hope you can assist me. Thank you very much!
Best Regards Bo