google-research/scenic

Dense Video Annotation Models

lrklein29 opened this issue · 0 comments

Is there a way to use the checkpoints available in the Vid2Seq repo with the corresponding PyTorch code? I see that the checkpoints in the VidChapters PyTorch repo are geared more towards chapters rather than the video captioning task. Are there PyTorch versions of the models trained on video captioning data?