Ask HN: Is GitHub Good for ML?

As part of my job I recently worked on an integration with GitHub and found out there are a lot of "open source" machine learning projects. Although GitHub is an incredible way to develop software I find there are some obvious shortcoming of using it for open source data science collaboration.

I'd love to hear your thoughts if you agree/disagree.

Main things that are missing from GitHub IMO:

- Uploading, viewing, & reviewing data+models

- CAN'T REVIEW NOTEBOOKS

- Some parts of the workflow like experiment tracking & labeling

(Disclaimer: I'm a developer at DagsHub)