hannabdul/etf4asr

Official repo for the paper "An Effective Training Framework for Light-Weight Automatic Speech Recognition Models" accepted at InterSpeech 2025.

13
/ 100
Experimental

This framework helps machine learning engineers and researchers optimize automatic speech recognition (ASR) models to be more efficient without sacrificing accuracy. It provides a structured approach for training ASR models using audio data and corresponding transcripts, resulting in 'light-weight' models that are faster and require fewer computational resources. It is designed for those who develop or deploy speech-to-text technologies.

No commits in the last 6 months.

Use this if you are a machine learning engineer working on speech recognition and need to train highly efficient, smaller ASR models for deployment on resource-constrained devices or for faster inference.

Not ideal if you are a business user looking for an off-the-shelf speech-to-text transcription service, as this tool is for developing the underlying ASR models.

speech-recognition-engineering machine-learning-optimization model-efficiency audio-processing natural-language-processing-development
No License Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 4 / 25
Maturity 7 / 25
Community 0 / 25

How are scores calculated?

Stars

8

Forks

Language

Lex

License

Last pushed

Aug 15, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/hannabdul/etf4asr"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.