Indicconformer Stt Ne Hybrid Ctc Rnnt Large
I
Indicconformer Stt Ne Hybrid Ctc Rnnt Large
Developed by ai4bharat
IndicConformer is a Conformer-based automatic speech recognition model with hybrid CTC-RNNT architecture, specifically optimized for Nepali language
Downloads 36
Release Time : 9/5/2024
Model Overview
This model adopts the Conformer-Large structure as the encoder, supporting transcription of Nepali speech content into text
Model Features
Hybrid Decoding Architecture
Supports both CTC and RNNT decoding methods, providing more flexible inference options
Large Capacity Model
Uses the 120-million-parameter Conformer-Large architecture with 512 feature dimensions to ensure recognition accuracy
Specialized Language Support
Speech recognition model specifically optimized for Nepali language
Model Capabilities
Nepali speech-to-text
16kHz mono audio processing
Use Cases
Speech Transcription
Nepali Meeting Minutes
Automatically transcribe Nepali meeting recordings into text records
Voice Assistant
Provide speech recognition capabilities for Nepali voice assistants
Featured Recommended AI Models
Š 2025AIbase