SPRING LAB HINDI-STREAMING
Automatic Speech Recognition (ASR) model for Hindi speech recognition, processing audio and transcribing spoken content into text.
About Model
Automatic Speech Recognition (ASR) model for Hindi speech recognition, developed using the Icefall toolkit with the Zipformer architecture. The model is trained on the Hindi dataset, which consists of approximately 500 hours of labelled speech. It is trained on 16 kHz audio, including naturally occurring code-mixed speech, enabling robust recognition of bilingual Indian speech patterns. The system is based on a 65M-parameter Zipformer-Medium encoder, paired with an RNN-T prediction network and joiner, forming a low-latency streaming ASR model with 16 encoder layers and a 512-dimensional representation.
SPRING LAB HINDI-STREAMING
Metadata
Attribution 4.0 International (CC BY- 4.0)
SPRING LAB IITM
Speech -to-text Conversion
PyTorch
Open
Science, Technology and Research
12/12/25 06:20:22
263.56 MB
jit_script_chunk_32_left_128.pt ( 263.53 MB )
To preview this file, you need to be a registered user. Please complete the registration process to gain access and continue viewing the content.
Activity Overview
0
14 
263.56 MB 
405
Tags
- ASR
- pytorch
- Hindi
- IITM
- spring_lab
- streaming
- MODELS
- zipformer
- Icefall-K2
License Control
Attribution 4.0 International (CC BY- 4.0)
Version Control
Version 1(263.56 MB)
- admin·5 month(s) ago
jit_script_chunk_32_left_128.pt
tokens.txt
More Models from Digital India BHASHINI Division
- 0
- 17
- 115.62 MB
106
- 0
- 13
- 2.08 GB
- 170
- 0
- 7
- 2.09 GB
- 112
- 0
- 3
- 3.52 GB
- 107
- 0
- 3
- 3.52 GB
- 100
- 0
- 4
- 3.52 GB
- 98
- 0
- 5
- 3.52 GB
- 138
- 0
- 3
- 3.52 GB
- 157
- 0
- 7
- 3.52 GB
- 156
- 0
- 5
- 3.52 GB
- 158
© 2026 - Copyright AIKosh. All rights reserved.