Retraining-Free Pruning Text-to-Speech Synthesis Model for Speaker Cloning

Authors: Ali Raheem Mandeel, Tamara Z. Fadhil, Mohammed Hamzah Alsalhi, Mohammed Salah Al-Radhi, Géza Németh

Email: {alimandeel, malradhi, nemeth}@tmit.bme.hu

Datasets

Sentence	Speech (Ground Truth)	Dense model	Lower_anchor	structured Wanda of sparsity 2/4	structured Wanda of sparsity 3/4	structured Wanda of sparsity 7/22	unstructured Wanda of sparsity 25%	unstructured Wanda of sparsity 50%	unstructured Wanda of sparsity 65%
01
02
03
04
05
06
07
08
09
10