The Kathbath-Sanskrit-Test-Unknown dataset is a comprehensive benchmark dataset designed to evaluate the performance of Automatic Speech Recognition (ASR) systems in Sanskrit. Featuring 1684 hours of labeled speech data across 12 Indian languages, it is tailored for general-domain testing scenarios. Submitted by Tahir Javed, this dataset is a valuable resource for advancing ASR technologies for Sanskrit and other regional Indian languages, facilitating the development of robust multilingual ASR systems.
© 2026 - Copyright AIKosh. All rights reserved.