fix: the exeption when topk is 1

inference.py

@@ -114,7 +114,7 @@ def main():
                 _logger.info('Predict: [{0}/{1}] Time {batch_time.val:.3f} ({batch_time.avg:.3f})'.format(
                     batch_idx, len(loader), batch_time=batch_time))
 
-    topk_ids = np.concatenate(topk_ids, axis=0).squeeze()
+    topk_ids = np.concatenate(topk_ids, axis=0)
 
     with open(os.path.join(args.output_dir, './topk_ids.csv'), 'w') as out_file:
         filenames = loader.dataset.filenames(basename=True)