TDS-CLIP: Temporal Difference Side Network for Image-to-Video Transfer Learning

Benchmark Model Rank Results
action-recognition-in-videos-on-somethingTDS-CLIP-ViT-L/14(8frames)#20Top-1 Accuracy: 73.4Top-5 Accuracy: 93.8
action-recognition-in-videos-on-something-1TDS-CLIP-ViT-L/14(8frames)#6Top 1 Accuracy: 63.0Top 5 Accuracy: 87.8