[graphGL]

> IMO, a release with open weights without the "source" is much better than the opposite, a release with open source and no trained weights.

Why not both?

[totetsu]

Isn’t everyone using data sets they don’t have rights to? Or at least not right to release.

[ebalit]

Both would be even better but it would be much smaller progress for the open source community in comparison to the release of the weights.