Maybe there is a bug in the Groupfree3D?

[Agito555]

I'm a beginner and not yet familiar with GroupFree3D. When I was looking at its code, I came across something quite confusing that the two-line code1 and code2 use 1-target to supervise its corresponding score when using focal loss. I'm not clear about the purpose of doing this, and I hope someone can help me figure it out.