Audio-to-Video on VGGSound (test)
150.26gFVDAVTok-A2V
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AVTok-A2VGen Type=AR, #Param Tokenizer=208.4M, #Param Generator=632.0M2026.06 | 150.26 | — | 1.317 | 0.143 | |
| TempoTokensGen Type=Diff, #Param Tokenizer=83.7M, #Param Generator=1.9B2026.06 | 786.61 | — | 1.359 | 0.132 |