How to write pytest tests for pytorch lightning custom callbacks?

[malfonsoarquimea]

Hi! I am currently trying to write a set of pytest tests for our codebase, and I have some questions.
My current approach to testing that the training is working as expected was to launch a training process within the pytest test using subprocess.run() with a given set of arguments (in the form of a config file) and then checking that the training process produces the expected files (checkpoints mostly) at the expected directories.
I think that this approach is not appropriate, so I would like to know If any of you could suggest a better way of testing the custom callbacks that I wrote that run when training

Thanks in advance!

[semaphore-egg]

Can we imitate the pl tests or something?