The kathbath_hard_punjabi dataset is a Punjabi Automatic Speech Recognition (ASR) benchmark dataset. Designed to evaluate ASR models for the Punjabi language, it includes challenging scenarios and diverse data from news and general domains. This dataset is an essential resource for researchers and developers working on Punjabi speech recognition, offering a robust foundation for building and benchmarking ASR systems. Submitted by AI4Bharat, it contributes to advancing speech technology for low-resource languages.
© 2026 - Copyright AIKosh. All rights reserved.