This ASR (Automatic Speech Recognition) model is designed for Assamese speech recognition. It processes 16,000 KHz mono WAV audio files and converts spoken Assamese into text. The model is based on a Conformer-Large architecture, featuring a 120M-parameter encoder and a hybrid CTC-RNNT decoder. It includes 17 conformer blocks, each with a model dimension of 512, ensuring efficient, accurate, and high-quality transcription of Assamese speech into text for various applications.
MIT
Sector Agnostic
AI4Bharat
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
Open
21/02/25 13:21:50
N.A.
0
© 2026 - Copyright AIKosh. All rights reserved.