BUT-FIT
/

diarizen-wavlm-large-s80-md

Voice Activity Detection

speaker-diarization

pyannote-audio-pipeline

Model card Files Files and versions

Jyhan003 commited on Jun 3

Commit

697d955

·

1 Parent(s): 5f93079

docs: update README

Files changed (1) hide show

README.md +2 -1

README.md CHANGED Viewed

@@ -14,7 +14,8 @@ tags:
 ## Overview
 This hub features the pre-trained model by [DiariZen](https://github.com/BUTSpeechFIT/DiariZen). The EEND component is built upon WavLM Large and Conformer layers. The model was trained on far-field, single-channel audio from a diverse set of public datasets, including AMI, AISHELL-4, AliMeeting, NOTSOFAR-1, MSDWild, DIHARD3, RAMC, and VoxConverse.
-Then structured pruning at 80% sparsity is applied. Smaller, faster, and better.
 ## Usage

 ## Overview
 This hub features the pre-trained model by [DiariZen](https://github.com/BUTSpeechFIT/DiariZen). The EEND component is built upon WavLM Large and Conformer layers. The model was trained on far-field, single-channel audio from a diverse set of public datasets, including AMI, AISHELL-4, AliMeeting, NOTSOFAR-1, MSDWild, DIHARD3, RAMC, and VoxConverse.
+Then structured pruning at 80% sparsity is applied. After pruning, the number of parameters in WavLM Large is reduced from **316.6M to 63.3M**, and the computational cost (MACs) decreases from **17.8G to 3.8G** per second.
 ## Usage