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

Question about ./datasets/LibriLight/test.txt #7

Closed xiaoli1996 closed 2 years ago

xiaoli1996 commented 2 years ago

In train.txt/val.txt/test.txt, what does “ 'cpc_km100': '-1' ” mean? Are there other available values for 'cpc_km100' apart from '-1'?

adampolyak commented 2 years ago

Unfortunately no.

We used LibriLight just for VQVAE training, therefore we didn't extract discrete codes for it for CPC or HuBERT.

xiaoli1996 commented 2 years ago

Did you mean that: if I want to use LibriLight just for vqvae training, I must set 'cpc_km100':-1, 'hubert':-1 and 'vqvae256':-1 in the train.txt? or just set one of them?

adampolyak commented 2 years ago

use cpc_km100

xiaoli1996 commented 2 years ago

Thank you very much!