Skip to main content

Week 5

Milestones

  • Feature addition
  • Text to voice integration model study
  • Python script for text to voice model

Screenshots / Videos

  • Model Study Text to Speech
  • API script Python script

Learnings

  • AI4Bharat Voice to text transcribe models AI4Bharat TTS is an open-source project by IIT Madras that provides state-of-the-art text-to-speech models for 13 Indian languages. The models are based on FastPitch and HiFiGAN, and they can generate high-quality speech with a variety of accents and speaking styles. The project also provides a set of tools and resources for developing and deploying TTS applications for Indian languages.

    Here are some of the key features of AI4Bharat TTS:

    State-of-the-art models for 13 Indian languages: The models are trained on large datasets of text and audio, and they can generate high-quality speech with a variety of accents and speaking styles. Open-source: The models and code are released under the Apache 2.0 license, so they can be freely used and modified by anyone. Easy to use: The project provides a set of tools and resources to make it easy to develop and deploy TTS applications for Indian languages.