Comparsion of German TTS models

The GlowTTS / Synestheisam model is the one I trained :slight_smile:

It doesn’t have the correct prosody (up/down pitch over time), but it was trained to be small and fast for Raspberry Pi’s. It’ll also be available in 2.5.8!


Wow, quality sounds really nice in the examples!
Looking forward to testing that in real life :star_struck:

1 Like