Starstruk Logo MarkStarstruk

v3.1

DigitalPhonetics/IMS-Toucanv3.1Jul 25, 2024by Flux9665

AI Summary

This release provides new checkpoints with stochastic prosody prediction that samples from distributions, additional IPA modifier support, more languages in pretraining, and overhauled language similarity prediction modules with visualization.

Key Highlights

  • Stochastic prosody prediction sampling from distributions
  • Added support for more IPA modifiers
  • Added more languages into pretraining
  • Overhauled language similarity prediction modules
  • Language similarity visualization

New Features

  • Stochastic prosody prediction (pitch, energy, durations)
  • More IPA modifiers for language coverage
  • Expanded language pretraining
  • Language similarity prediction overhaul

Full Release Notes

## What's Changed

This release provides new checkpoints and improves some aspects of the previous release that were not included due to time constraints.  For more information on the universal TTS model for 7000 languages, please refer to the previous release v3.0

- Prosody prediction in terms of pitch, energy and durations are now stochastic and sample from a distribution instead of assuming a one-to-one mapping. 
- Added support for more IPA modifiers to cover more languages
- Added more languages into the pretraining 
- Overhauled language similarity prediction modules and visualization

**Full Changelog**: https://github.com/DigitalPhonetics/IMS-Toucan/compare/v3.0...v3.1