GigaAM Multilingual: Foundation Model for Underrepresented Languages
This paper develops a multilingual speech recognition model for underrepresented languages, addressing the issue of data scarcity and uneven performance, and shows promising results in controlled comparisons.