search

facebookresearch / speech-resynthesis

An official reimplementation of the method described in the INTERSPEECH 2021 paper - Speech Resynthesis from Discrete Disentangled Self-Supervised Representations.
Other
390 stars 54 forks source link

speaker information #21

Open speechless-z opened 1 year ago

speechless-z commented 1 year ago

Will the code used for extracting speaker information in the paper not be provided?

speechless-z commented 1 year ago

Hello, I carefully reviewed your code and tried to replicate it. I have some doubts regarding the speaker embedding extraction part. The code initializes a matrix randomly and adjusts its parameters during training. Does this mean that a pre-trained speaker model mentioned in the model graph is not necessary？？