No comments yet.
One thing we kept running into is users being challenged by setting up the cloud environment necessary, to push their data to AWS S3, GCS (which we support), and other DVC remotes. Since our focus is on community projects, it was clear that this is a lot of friction that we don't want, as it gets in the way of users actually working on data science.
To fix this, we built DAGsHub Storage. It's the easiest DVC remote you can set up, and we made the process of setting it up as close as possible to the process of setting up a Git remote. That means that you don't need a credit card, define storage access permissions, IAM, etc. Just type in `dvc remote add origin [https://dagshub.com/<user>/<project>.dvc](https://dagshub.com/<user>/<project>.dvc)`.
Since our goal is to spur community projects and collaboration, DAGsHub storage is free for open source projects, and we also gave some free storage to private projects, because, heck, why not.
After you push your data to DAGsHub Storage, you'll be able to view it, compare versions and share it with collaborators easily, as well as control team access. This means that you can contribute data to projects, not just code. I hope this makes ML work better and promote Open Source Data Science.
Unlike GitLFS, we're offering significantly more storage, basing it on the much more data science oriented DVC, and actually letting you browse, compare and share the data files and artifacts via the UI.
As usual, I'd love to hear your feedback!
https://dagshub.com/blog/dagshub-storage-zero-configuration-dataset-model-hosting
