g1910/HyperNetworks

Reasoning behind parameter initialization?

adam-katona opened this issue · 1 comments

adam-katona commented

Hi,

I was wondering what is the reason the hypernetwork parameters are initialized like this:
torch.fmod(torch.randn(dim),2)

By the way, thanks for the repo, it is awesome.

g1910 commented

There is no specific reasoning behind doing this parameter initialization. You can do it any way you like (for example, using xavier initialization). Hope this helps! :)