[D] How do you get enough patience to train-debug-train models?

Train a big model only for it to get errors. I don't multi-task because focusing on one thing gets it completed faster.

But it is so annoying. I just wait forever for it to get completed while all my other tasks are paused and when I get the results, it is all errors!

Eugh brother

big debug errors faster machinelearning results tasks train