Using a big file (e.g ML model) as part of CI, in a fast way

I have an ML related project, which as a part of the CI process needs an existing “big” (multiple GB) model pickle file, to be present at the running container
I don’t want to download it every run, is there a way to make it “available” to my runs and store it somewhere inside gitlab? (without pushing the gigantic file to git of course)