alexandonian / contrastive-feature-loss

According to the equation (5) in paper, which says that "gradients are prevented from flowing through the negatives". Then for the PatchNCELoss function in

contrastive-feature-loss/models/networks/nce.py

Line 30 in 788ac92

if not self.opt.gradient_flows_to_negative_nce:

, should we detach feat_k after the calculation of pos logits?

Problem about the stop gradient operation for negative pairs.