Audio-to-Video generation on VGGSound (test)
402.385FVDDiffusion Latent Aligner
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Diffusion Latent Alignervariant=full, denoising_steps=252024.02 | 402.385 | 34.764 | 52.2 | |
| Diffusion Latent Alignervariant=vanilla, denoising_steps=252024.02 | 417.398 | 36.262 | 51.8 | |
| TempoToken2024.02 | 1,866.285 | 389.096 | 42.3 |