Great Work on DistillKitPlus + Licensing Consideration

[Crystalcareai]

Hi there,

Lucas here—I lead the Labs Research team at Arcee and was one of the primary creators of DistillKit. First off, I want to say how impressed I am with DistillKitPlus. You’ve done a fantastic job! Your changes are thoughtful, especially given that our focus shifted away from this project before reaching its full potential. Seeing the torch carried forward is outstanding.

That said, I’d like to flag a small licensing consideration. Some of the code in DistillKitPlus appears to be carried over from our original DistillKit work, which was initially licensed under AGPL-3.0. The change to Apache-2.0 in your project introduces a bit of a mismatch since AGPL-3.0 has stricter requirements around derivative works. While this isn’t intended as a criticism or to cause concern, I wanted to highlight it as something to keep in mind for future derivative works.

At Arcee, we’re deeply committed to fostering collaboration in the open-source ML community. To make things simpler and more aligned with projects like yours, we’ve actually decided to relicense DistillKit under Apache-2.0. You can view the update here: Commit Link.

In addition, we’d love to invite you to contribute some of your improvements to the main DistillKit repo. If you’re open to it, submitting a PR would be an excellent way to collaborate and ensure your enhancements benefit the broader ML community.

Thank you again for your incredible work. If you ever want to discuss licensing, ideas, or collaboration, feel free to reach out to me at lucas@arcee.ai or on X (@lucasatkins7).

Thanks again,
Lucas

[agokrani]

@Crystalcareai thank you so much for your kind words and for appreciating my work by making license less restrictive. 🙂 You guys have been amazing contributors to the opensource ecosystem.

[maziyarpanahi]

Thank you @Crystalcareai and Arcee for this classy move! And thank you @agokrani for sharing your contributions.