Method validate_classifier

src/worker.py:1628–1668

(self,model, generator, generator_mapping, generator_synthesis, epoch, GAN_test, setting)

    # validate GAN_train or GAN_test classifier using generated or training dataset
    # -----------------------------------------------------------------------------
    def validate_classifier(self,model, generator, generator_mapping, generator_synthesis, epoch, GAN_test, setting):
        model.eval()
        valid_top1_acc, valid_top5_acc, valid_loss = misc.AverageMeter(), misc.AverageMeter(), misc.AverageMeter()
        for i, (images, labels) in enumerate(self.train_dataloader):
            if GAN_test:
                images, labels, _, _, _, _, _ = sample.generate_images(z_prior=self.MODEL.z_prior,
                                                                 truncation_factor=self.RUN.truncation_factor,
                                                                 batch_size=self.OPTIMIZATION.batch_size,
                                                                 z_dim=self.MODEL.z_dim,
                                                                 num_classes=self.DATA.num_classes,
                                                                 y_sampler="totally_random",
                                                                 radius="N/A",
                                                                 generator=generator,
                                                                 discriminator=self.Dis,
                                                                 is_train=False,
                                                                 LOSS=self.LOSS,
                                                                 RUN=self.RUN,
                                                                 MODEL=self.MODEL,
                                                                 device=self.local_rank,
                                                                 is_stylegan=self.is_stylegan,
                                                                 generator_mapping=generator_mapping,
                                                                 generator_synthesis=generator_synthesis,
                                                                 style_mixing_p=0.0,
                                                                 stylegan_update_emas=False,
                                                                 cal_trsp_cost=False)
            else:
                images, labels = images.to(self.local_rank), labels.to(self.local_rank)

            output = model(images)
            ce_loss = self.ce_loss(output, labels)

            valid_acc1, valid_acc5 = misc.accuracy(output.data, labels, topk=(1, 5))

            valid_loss.update(ce_loss.item(), images.size(0))
            valid_top1_acc.update(valid_acc1.item(), images.size(0))
            valid_top5_acc.update(valid_acc5.item(), images.size(0))

        if self.local_rank == 0:
            self.logger.info("Top 1-acc {top1.val:.4f} ({top1.avg:.4f})\t"
                             "Top 5-acc {top5.val:.4f} ({top5.avg:.4f})".format(top1=valid_top1_acc, top5=valid_top5_acc))
        return valid_top1_acc.avg, valid_top5_acc.avg, valid_loss.avg