Skip to content

Latest commit

 

History

History
8 lines (5 loc) · 466 Bytes

README.md

File metadata and controls

8 lines (5 loc) · 466 Bytes

Text 2 Speech Engine Web App

This repo is an implementation of Tacotron 2 wrapped underneath a web application. I manage to generate a David Attenbourgh TTS model from training over 20 hours of labeld audio transcript data

Current Goals

  • Integrate maximizing mutual information into the loss function
  • Create APRAbet converter to break down text representation into lower dimensional phenome representation