Both the jet data (and star for some reason?) fail every now and then due to changes in the version of the libraries or even different workers running the workflow.
These should be fixed. Sometimes it's just the random 1/-1 when breaking down the covmat into systematics. It doesn't have a phenomenological effect but it is a lot of changes making the repository unnecessarily big. With a bit of care it can be made deterministic.
Both the jet data (and star for some reason?) fail every now and then due to changes in the version of the libraries or even different workers running the workflow.
These should be fixed. Sometimes it's just the random 1/-1 when breaking down the covmat into systematics. It doesn't have a phenomenological effect but it is a lot of changes making the repository unnecessarily big. With a bit of care it can be made deterministic.