The Automatic Speech Recognition (ASR) model is designed to transcribe spoken Gujarati into text from 16kHz mono-channel audio files. It utilizes a Conformer-Large architecture with 120 million parameters, ensuring robust and efficient speech recognition. The model employs a hybrid CTC-RNNT decoder for enhanced transcription accuracy. With 17 conformer blocks and a model dimension of 512, it effectively captures linguistic and acoustic nuances, making it suitable for various speech-to-text applications in Gujarati.
MIT
Sector Agnostic
AI4Bharat
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
Open
21/02/25 13:21:46
N.A.
0
© 2026 - Copyright AIKosh. All rights reserved.