Turns out Linux is killing the kernel after it’s taken almost the entirety of the memory. Tracing the culprit in the lab codes, turns out I forgot to turn off loading the gigantic dataset LMAO.
Lighter dataset is now used and the PyTorch code is executing just fine without the Python kernel dying anymore.
This site will hold short thoughts, daily observations, technical discoveries, project progress, and records that do not require a full article.