Parameter-efficient transfer learning of pre-trained Transformer models for speaker verification using adapters.