Spaces:

---

nari-labs

/

Dia-1.6B

like 647

Running on Zero

Hi! I am on the gradio team and I am developing a new component to make it easier for users to specify a dialogue between multiple speakers without needing to know the exact chat template used by the underlying model.

Features:

• Nice UI for adding/removing lines from a dialogue
• The component automatically formats the dialogue into the format expected by the model
• An autocomplete menu for the different supported intonations and emotions appears when you type colon (:)

LGTM!