Universal Pooling Method of Multi-layer Features from Pretrained Models for Speaker Verification

Open in new window