Improved spell corrector algorithm and deepspeech2 model for enhancing end-to-end Gujarati language ASR performance
Automatic Speech Recognition (ASR) is the process of converting auditory signals into text representations of spoken words.In recent years, advancements in deep learning algorithms have resulted in the development of intricate architectures that considerably enhance the efficacy of End-to-End (E2E) ASR systems.Obtaining significant quantities of tr