MeMViT: Memory-Augmented Multiscale Vision Transformer for Efficient Long-Term Video Recognition

Open paper
Benchmark Model Rank Results
action-anticipation-on-epic-kitchens-100MeMViT-24#4Recall@5: 17.7
action-recognition-on-ava-v2-2MeMViT-24#16mAP: 35.4
action-recognition-on-epic-kitchens-100MeMViT-24#10Action@1: 48.4Verb@1: 71.4Noun@1: 60.3

Rank counts only results with a code link.