MMAudio V2

This model generates audio to accompany a video. You can provide a text prompt to guide the sound, or leave the prompt empty to let the model infer suitable audio from the video alone.