This repository presents an evaluation framework for speech-to-speech (S2S) models, following the methodology described in the EmphAsses paper (de Seyssel et al., 2023). - View it on GitHub