adirik / seamless-expressive

Multilingual speech translation that preserves original vocal style and prosody

  • Public
  • 1.2K runs
  • GitHub
  • Paper
  • License

No versions have been pushed to this model yet.