Source of the dataset is AudioSet(Gemmeke et al., 2017) from Google Research. Music Captions are generated with MU-LLaMA, Video Captions with SwinBERT, MSI and Fusion with Mixtral from both information.
Source of the dataset is AudioSet(Gemmeke et al., 2017) from Google Research. Music Captions are generated with MU-LLaMA, Video Captions with SwinBERT, MSI and Fusion with Mixtral from both information.