Danish Text-to-SpeechDanish Text-to-SpeechOverview
A Danish text-to-speech adapter built on top of sesame/csm-1b. The work here is making Danish sound right: cleaner pronunciation, better pacing, and fewer “weird” drops in longer sentences.
Trained as a LoRA adapter on a mixed Danish dataset (public corpora + a private extension). Data is filtered/normalized and text is cleaned to keep training stable and output consistent. The setup also supports two voice presets controlled from the prompt.
Packaged with a simple demo and curated audio samples, so it’s easy to test quickly and share results. Released under Apache-2.0 and requires access to the base model.