[dheera]

What if there are two successors of YOLOv4 that both inherit large parts of YOLOv4, make different sets of changes, achieving different sets of improvements?

Do we call them YOLOv5.0 and YOLOv5.1? How do we nomenclate the graph? What if another one inherits from both YOLOv5.0 and YOLOv5.1? Is it YOLOv6 or YOLOv(5.0,5.1).0?

[renewiltord]

A problem for when it happens. In the past, the first guys to publish theirs to the public get the first one. The other guys can just call theirs Yolo6 or miniYolo or whatever. It's not a big deal. We don't need a system for conflict resolution if no conflict is likely.

There is no controversy. Let it lie.

[CJefferson]

This type of thing does happen in academia. There is a family of algorithms in constraint programming called AC, and there is AC versions 1,2,3,4,5,6,7,3.1,3.2,3.3,2000,3.1* any several other variants.

These algorithms are (mostly) written by different people. People build on each others work in academia all the timr.

[yeldarb]

In the business world we have trademarks for this.

As models become brand names and franchises I wonder if academics will start trademarking them like many big open source projects do.

[dheera]

Although that's true, I don't think it's necessary to trademark in this case. Nobody is losing marketing potential or profits here -- it's just an issue of how to deal with nomenclature when two separate people are working on different forked improvements of the same thing without confusing the community.

[paulryanrogers]

This is where trademarks and foundations may need to be used.