What is emb_scale used for?

[Ocelot7777]

id_head = self.emb_scale * F.normalize(id_head) So why do you use emb_scale?

[faruknane]

@Ocelot7777 I have to ask the same question. It throws an error at this line when opt.nID is 1. So is there a way that we can avoid using emb_scale and what does it do?

[zengjie617789]

It is hardly to train a network on one image~~~

[nmll]

I also have the question about that why to use this operation about self.emb_scale? id_head = self.emb_scale * F.normalize(id_head)