Class Token and Knowledge Distillation for Multi-head Self-Attention Speaker Verification Systems

Open in new window