okay, that is a step in the right direction. Unfortunately, now the script is not stored.
Maybe I don't use clearml as intended. As a first try, I run the code locally, then I clone it and run it on a worker.
With the above change, there is no code stored: "Note: This experiment seems to not contain any executable code. If enqueued to run, it might produce no results."

Hi HighCentipede85 ,

You can pass the git repo detection with auto_connect_frameworks={"detect_repository": False} :
task = Task.init(project_name="Your project name", task_name="Your task name", auto_connect_frameworks={"detect_repository": False})Notice: This will leave empty all the sections for the execution part.

Sure, I have opened https://github.com/allegroai/clearml/issues/340 .

or do you mean this by your note? i.e. leaving the execution parts empty.

yep 🙂

Currently, when ClearML detect a .git file, it will store your running script as part of your git repo. The workaround to store the whole script as a standalone is just like you did (or to run it outside of the repo).
We currently don’t have such an option to store the script as a standalone, but this could be a useful feature.

Can you add a new  https://github.com/allegroai/clearml/issues  issue with this request? Just so it won’t get lost

or do you mean this by your note? i.e. leaving the execution parts empty.