INNER CODE UNIT · Python

loss

KaiyangZhou/pytorch-center-loss · main.py:122

        loss = loss_xent + loss_cent
        optimizer_model.zero_grad()
        optimizer_centloss.zero_grad()
        loss.backward()
        optimizer_model.step()
        # by doing so, weight_cent would not impact on the learning of centers
        for param in criterion_cent.parameters():
            param.grad.data *= (1. / args.weight_cent)
        optimizer_centloss.step()
        
        losses.update(loss.item(), labels.size(0))
        xent_losses.update(loss_xent.item(), labels.size(0))
        cent_losses.update(loss_cent.item(), labels.size(0))

        if args.plot:
            if use_gpu:
                all_features.append(features.data.cpu().numpy())
                all_labels.append(labels.data.cpu().numpy())

View source record →