Result Details
AN ATTENTION-BASED BACKEND ALLOWING EFFICIENT FINE-TUNING OF TRANSFORMER MODELS FOR SPEAKER VERIFICATION
Created: 2024
Type
software
Language
English
Authors
Peng Junyi, DCGM (FIT)
Plchot Oldřich, Ing., Ph.D., DCGM (FIT)
Stafylakis Themos
Mošner Ladislav, Ing., Ph.D., DCGM (FIT)
Burget Lukáš, doc. Ing., Ph.D., DCGM (FIT)
Černocký Jan, prof. Dr. Ing., DCGM (FIT)
Plchot Oldřich, Ing., Ph.D., DCGM (FIT)
Stafylakis Themos
Mošner Ladislav, Ing., Ph.D., DCGM (FIT)
Burget Lukáš, doc. Ing., Ph.D., DCGM (FIT)
Černocký Jan, prof. Dr. Ing., DCGM (FIT)
Description
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.
Keywords
speech recognition, target speech processing, speaker, verification, anti-spoofing, language identification
URL
License
In order to use the result by another entity, it is always necessary to acquire a license
License Fee
The licensor does not require a license fee for the result
Projects
Linguistics, Artificial Intelligence and Language and Speech Technologies: from Research to Applications, EU, MEZISEKTOROVÁ SPOLUPRÁCE, EH23_020/0008518, start: 2025-01-01, end: 2028-12-31, running
Research groups
Speech Data Mining Research Group BUT Speech@FIT (RG SPEECH)
Departments