Detail aplikovaného výsledku

AN ATTENTION-BASED BACKEND ALLOWING EFFICIENT FINE-TUNING OF TRANSFORMER MODELS FOR SPEAKER VERIFICATION

PENG, J.; PLCHOT, O.; STAFYLAKIS, T.; MOŠNER, L.; BURGET, L.; ČERNOCKÝ, J.

Originální název

AN ATTENTION-BASED BACKEND ALLOWING EFFICIENT FINE-TUNING OF TRANSFORMER MODELS FOR SPEAKER VERIFICATION

Anglický název

AN ATTENTION-BASED BACKEND ALLOWING EFFICIENT FINE-TUNING OF TRANSFORMER MODELS FOR SPEAKER VERIFICATION

Druh

Software

Abstrakt

MHFA is an advanced speaker extractor back-end applied to layerwise representations of speech foundation models. It has been successfully used not only in speaker verification, but also in tasks and challenges such as anti-spoofing, language identification, and target speech processing.

Abstrakt anglicky

MHFA is an advanced speaker extractor back-end applied to layerwise representations of speech foundation models. It has been successfully used not only in speaker verification, but also in tasks and challenges such as anti-spoofing, language identification, and target speech processing.

Klíčová slova

speech recognition, target speech processing, speaker, verification, anti-spoofing, language identification

Klíčová slova anglicky

speech recognition, target speech processing, speaker, verification, anti-spoofing, language identification

Licenční poplatek

K využití výsledku jiným subjektem je vždy nutné nabytí licence

www

https://github.com/JunyiPeng00/SLT22_MultiHead-Factorized-Attentive-Pooling