Pol-Buitrago / SynthAVSR

关于

This repository contains the development of SynthAVSR, the first Audiovisual Speech Recognition (AVSR) system tailored for the Spanish and Catalan languages. Based on the AV-HuBERT (Audio-Visual Hidden Unit BERT) model, SynthAVSR leverages synthetic audiovisual data to bridge the gap in speech recognition technology for these languages.