SpeechBrain: A General-Purpose Speech Toolkit
Titouan Parcollet
(1)
,
Mirco Ravanelli
(2)
,
Peter Plantinga
(3)
,
Aku Rouhe
(4)
,
Samuele Cornell
(5)
,
Loren Lugosch
(6, 2)
,
Cem Subakan
(2)
,
Nauman Dawalatabad
(7)
,
Abdelwahab Heba
(8)
,
Jianyuan Zhong
(2)
,
Ju-Chieh Chou
(9)
,
Sung-Lin Yeh
(10)
,
Szu-Wei Fu
(11)
,
Chien-Feng Liao
(11)
,
Elena Rastorgueva
(12)
,
François Grondin
(13)
,
William Aris
(14)
,
Hwidong Na
(15)
,
Yan Gao
(16)
,
Renato de Mori
(1, 6)
,
Yoshua Bengio
(2)
1
LIA -
Laboratoire Informatique d'Avignon
2 MILA - Montreal Institute for Learning Algorithms [Montréal]
3 OSU - The Ohio State University [Columbus]
4 Aalto University
5 UNIVPM - Polytechnic University of Marche / Università Politecnica delle Marche
6 McGill University = Université McGill [Montréal, Canada]
7 IIT Madras - Indian Institute of Technology Madras
8 IRIT-SAMoVA - Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio
9 TTIC - Toyota Technological Institute at Chicago [Chicago]
10 Edin. - University of Edinburgh
11 Academia Sinica
12 NVIDIA - NVIDIA
13 UdeS - Université de Sherbrooke
14 UdeS - Département d'informatique [Sherbrooke]
15 SAIT - Samsung Advanced Institute of Technology
16 CAM - University of Cambridge [UK]
2 MILA - Montreal Institute for Learning Algorithms [Montréal]
3 OSU - The Ohio State University [Columbus]
4 Aalto University
5 UNIVPM - Polytechnic University of Marche / Università Politecnica delle Marche
6 McGill University = Université McGill [Montréal, Canada]
7 IIT Madras - Indian Institute of Technology Madras
8 IRIT-SAMoVA - Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio
9 TTIC - Toyota Technological Institute at Chicago [Chicago]
10 Edin. - University of Edinburgh
11 Academia Sinica
12 NVIDIA - NVIDIA
13 UdeS - Université de Sherbrooke
14 UdeS - Département d'informatique [Sherbrooke]
15 SAIT - Samsung Advanced Institute of Technology
16 CAM - University of Cambridge [UK]
Titouan Parcollet
- Fonction : Auteur
- PersonId : 174514
- IdHAL : titouan-parcollet
- ORCID : 0000-0003-0672-1346
Abdelwahab Heba
- Fonction : Auteur
- PersonId : 1129713
- IdRef : 221309241
Renato de Mori
- Fonction : Auteur
- PersonId : 981954
Résumé
SpeechBrain is an open-source and all-in-one speech toolkit. It is designed to facilitate the research and development of neural speech processing technologies by being simple, flexible, user-friendly, and well-documented. This paper describes the core architecture designed to support several tasks of common interest, allowing users to naturally conceive, compare and share novel speech processing pipelines. SpeechBrain achieves competitive or state-of-the-art performance in a wide range of speech benchmarks. It also provides training recipes, pretrained models, and inference scripts for popular speech datasets, as well as tutorials which allow anyone with basic Python proficiency to familiarize themselves with speech technologies.