The error runs for a while then stops automatically

[Jatbi]

I ran everything very well. But every once in a while it will automatically stop. I'm using tmux so it can run in the background.

then I have to ctrl + c it to stop everything and kill -9 it, and starting it again works but normally. But every time it takes 5 minutes or 10 minutes

i am using ubuntu 22.

NVIDIA-SMI 550.54.15 Driver Version: 550.54.15

CUDA Version: 12.1

I've researched but haven't seen anyone like that. but before that everything worked very well.

[Jatbi]

When I checked the progress it was still running but very slowly. I had to stop it and run it again for everything to work as usual.
I checked via htop that everything is not running beyond the server configuration I have.