iliya-malecki
commented
3 years ago okay, now I'm thinking it has to do with the ability to reuse the same compute clusters without the need to spin up new ones and send the data to them. If that's true, please clarify that in the documentation, that took a long time to figure out on my own. And yes, I'm still not sure if I'm correct.
I'm quite new to azure, so maybe I just don't understand something, but the first thing I did trying to understand this template was to track what calls what and where. And I found that scripts in the ci pipeline get the whole diabetes_regression folder as a source. Is it intentional or is it a mishap? Isnt it more logical to make a "CI" folder in "diabetes_regression", put only CI scripts there and do the os.path.join(e.sources_directory, "[script_step_subdirectory_name]")?