GitLab

Summary:
The PR adds a tutorial that demonstrates how to use pre-trained `TorchAudio-SQUIM` pipelines to estimate objective and subjective metric scores (PESQ, STOI, Si-SDR, MOS).

Pull Request resolved: https://github.com/pytorch/audio/pull/3279

Reviewed By: hwangjeff

Differential Revision: D45415404

Pulled By: nateanl

fbshipit-source-id: abcaeadcca0eabc2dca53b607eac6257a701c903