Skip to content

Open-Speech-EkStep/audio-to-speech-pipeline

Repository files navigation

Intelligent Data Pipeline

It is a part of Vakyansh's recipes to build state of the art Speech Recogniition Model

Pipelines to transform raw data and prepare data for model training. They clean, process and balance data for model training

All Contributors

Contributor Covenant License: MIT Python 3.8

CircleCI

Coverage Code Style Conventional Commits Gitter

If you like Vakyansh, ⭐ the project to support its development!_

Key Highlights

  • Automated: Vakyansh's pipelines are fully automated and no human intervention is required to process the audio pipelines. It uses pre trained models to identify language, gender and speakers in the audio.
  • Scalable: Process thousands of hours of data in 2-3 days leveraging K8s capabilities. Allows faster time to market.
  • Versatile: Vakyansh's pipelines can run on any VM's or Kubernetes cluster on any cloud. For scalability it is recommended to run on Kubernetes
  • Quality: Vakyansh lays strong emphasis on code quality and high test coverage to ensure its seamless use and continued maintenance.

Documentation

Please follow the documentation for Intelligent Data Pipeline here

You can improve it by sending pull requests to the documentation repository.

License

MIT License

About

This will hold the data pipeline to convert raw audio data to speech which will act as input dataset for speech-to-text pipeline

Resources

License

Code of conduct

Stars

Watchers

Forks

Packages

No packages published

Languages