paper-with-me

Action Recognition In Videos 벤치마크

Action Recognition In Videos on Kinetics-400

9개 결과 · ⬇ CSV · JSON

Top-1 Accuracy

81.7 82.9 84.1 85.3 86.5 2021-09 2026-09 ActionCLIP (ViT-B/16) — 83.8 (2021-09-17) ActionCLIP (ViT-B/16) — 83.8 (2021-09-17) ActionCLIP (ViT-B/16) — 83.8 (2021-09-17) Florence — 86.5 (2021-11-22) Florence — 86.5 (2021-11-22) Florence — 86.5 (2021-11-22) Frozen Backbone, SwinV2-G-ext22K (Video-Swin) — 81.7 (2022-11-03) Frozen Backbone, SwinV2-G-ext22K (Video-Swin) — 81.7 (2022-11-03) Frozen Backbone, SwinV2-G-ext22K (Video-Swin) — 81.7 (2022-11-03) ActionCLIP (ViT-B/16) — 83.8 (2021-09-17) Florence — 86.5 (2021-11-22)
RankModel Top-1 AccuracyTop-5 Accuracy PaperCodeYear
1 Florence 86.597.3 Florence: A New Foundation Model for Computer Vision microsoft/unicl · MindCode-4/code-3 2021
2 ActionCLIP (ViT-B/16) 83.8– ActionCLIP: A New Paradigm for Video Action Recognition towhee-io/towhee · sallymmx/actionclip 2021
3 Frozen Backbone, SwinV2-G-ext22K (Video-Swin) 81.7– Could Giant Pretrained Image Models Extract Universal Representations? 2022
4 Florence 86.597.3 Florence: A New Foundation Model for Computer Vision microsoft/unicl · MindCode-4/code-3 2021
5 ActionCLIP (ViT-B/16) 83.8– ActionCLIP: A New Paradigm for Video Action Recognition towhee-io/towhee · sallymmx/actionclip 2021
6 Frozen Backbone, SwinV2-G-ext22K (Video-Swin) 81.7– Could Giant Pretrained Image Models Extract Universal Representations? 2022
7 Florence 86.597.3 Florence: A New Foundation Model for Computer Vision microsoft/unicl · MindCode-4/code-3 2021
8 ActionCLIP (ViT-B/16) 83.8– ActionCLIP: A New Paradigm for Video Action Recognition towhee-io/towhee · sallymmx/actionclip 2021
9 Frozen Backbone, SwinV2-G-ext22K (Video-Swin) 81.7– Could Giant Pretrained Image Models Extract Universal Representations? 2022
1–9 / 9 페이지당 10 20 50 100