TADA

About

Hume AI offers TADA, an open-source speech-language model for perfect text and audio synchronization. TADA, or Text-Acoustic Dual Alignment, achieves a one-to-one correspondence between text and speech tokens, creating a unified and uninterrupted flow. This model produces audio five times faster than typical LLM-based text-to-speech, effectively eradicating skipped words and content hallucinations, as proven through over 1000 evaluations.