INNER CODE UNIT · Python

output1

adobe/antialiased-cnns · main.py:568

                output1 = model(input[:,:,off1[0]:off1[0]+224,off1[1]:off1[1]+224])

                cur_agree = agreement(output0, output1).type(torch.FloatTensor).to(output0.device)

                # measure agreement and record
                consist.update(cur_agree.item(), input.size(0))

                # measure elapsed time
                batch_time.update(time.time() - end)
                end = time.time()

                if i % args.print_freq == 0:
                    print('Ep [{0}/{1}]:\t'
                          'Test: [{2}/{3}]\t'
                          'Time {batch_time.val:.3f} ({batch_time.avg:.3f})\t'
                          'Consist {consist.val:.4f} ({consist.avg:.4f})\t'.format(
                           ep, args.epochs_shift, i, len(val_loader), batch_time=batch_time, consist=consist))

View source record →