I’ve abandoned Github to store my own code and have, instead, opted to simply dump it into a shared read-only Google Drive folder.
Too much trouble, and I don’t really need deep source control, even though Google offers it with its Google Drive, for free.
What really set me onto this choice was the apparent bias Atlassian Bitbucket has for retaining code-like material rather than datasets, and their 2 Gb ceiling. I was willing to live with that, but, then, fell into the difficulty of trying to expunge some largish datasets and having to prune them from the Github history, all doing this from Windows Sourcetree, or, rather, the command line side support it offers, which is hobbled.
I use Sourcetree to grab and keep up with others Github repositories.