Benchmarking Single-Factor Physical Video-to-Audio Generation
Open paper
Benchmark
Model
Rank
Results
video-to-sound-generation-on-vggsound-v2a
MMAudio-Phys (w/ Caption) (extra data)
–
ImageBind Score: 34.89
De-Sync: 0.405
KL (PANNs): 1.37
Rank counts only results with a code link.