Dia 1.6B 0626

JSON →
nari-labs audio
audiotext

A 1.6 billion parameter text-to-speech model from Nari Labs for expressive speech synthesis.