A 3-billion parameter base model for audio generation by Boson AI, capable of producing high-fidelity speech and sounds.
No resource links recorded.