MACCIF-TDNN: Multi aspect aggregation of channel and context interdependence features in TDNN-based speaker verification

Installation

MACCIF-TDNN is implemented based on Speechbrain https://github.com/speechbrain/speechbrain. Make sure SpeechBrain is installed correctly.

(1) Add MACCIF.py, MultiHeadAttentionPoolingC.py, PositionEncode.py into speechbrain/lobes/models/ folder;

(2) Modify speechbrain/recipes/VoxCeleb/SpeakerRec/hparams/train_ecapa_tdnn.yaml

Refer to the details in the paper.

Name		Name	Last commit message	Last commit date
Latest commit History 12 Commits
MACCIF.py		MACCIF.py
MultiHeadAttentionPoolingC.py		MultiHeadAttentionPoolingC.py
PositionEncode.py		PositionEncode.py
README.md		README.md
img.png		img.png
train_ecapa_tdnn.yaml		train_ecapa_tdnn.yaml