Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Yingzhi, Alhmoud, Anas, Alqurishi, Muhammad
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2412.13788
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913617062920192
author Wang, Yingzhi
Alhmoud, Anas
Alqurishi, Muhammad
author_facet Wang, Yingzhi
Alhmoud, Anas
Alqurishi, Muhammad
contents In recent years, the enhanced capabilities of ASR models and the emergence of multi-dialect datasets have increasingly pushed Arabic ASR model development toward an all-dialect-in-one direction. This trend highlights the need for benchmarking studies that evaluate model performance on multiple dialects, providing the community with insights into models' generalization capabilities. In this paper, we introduce Open Universal Arabic ASR Leaderboard, a continuous benchmark project for open-source general Arabic ASR models across various multi-dialect datasets. We also provide a comprehensive analysis of the model's robustness, speaker adaptation, inference efficiency, and memory consumption. This work aims to offer the Arabic ASR community a reference for models' general performance and also establish a common evaluation framework for multi-dialectal Arabic ASR models.
format Preprint
id arxiv_https___arxiv_org_abs_2412_13788
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Open Universal Arabic ASR Leaderboard
Wang, Yingzhi
Alhmoud, Anas
Alqurishi, Muhammad
Computation and Language
In recent years, the enhanced capabilities of ASR models and the emergence of multi-dialect datasets have increasingly pushed Arabic ASR model development toward an all-dialect-in-one direction. This trend highlights the need for benchmarking studies that evaluate model performance on multiple dialects, providing the community with insights into models' generalization capabilities. In this paper, we introduce Open Universal Arabic ASR Leaderboard, a continuous benchmark project for open-source general Arabic ASR models across various multi-dialect datasets. We also provide a comprehensive analysis of the model's robustness, speaker adaptation, inference efficiency, and memory consumption. This work aims to offer the Arabic ASR community a reference for models' general performance and also establish a common evaluation framework for multi-dialectal Arabic ASR models.
title Open Universal Arabic ASR Leaderboard
topic Computation and Language
url https://arxiv.org/abs/2412.13788