Skip to content

Files

Latest commit

b98a1bd · Feb 10, 2020

History

History
This branch is up to date with facebookresearch/CPC_audio:main.

cpc

Repository's architecture

train.py : main script

dataset.py : defintion of the Librispeech dataset format

model.py : Basic encoders and AR models

feature_loader.py: different tools to load and save a CPC model.

transformers.py: an implementation of transformers

unit_tests.py : unit tests

criterion/: definition of the training criterions. Three criterion are currently available: CPC (unsupervised), speaker classification and phone classification.

eval/: evaluation scripts.

utils/: system utilities and misc.