Temporal Sentence Grounding on Charades-STA

Metric: R5@0.5 (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	R5@0.5▼	Extra Data	Paper	Date↕	Code
1	DeCafNet	91.53	No	DeCafNet: Delegate and Conquer for Efficient Tem...	2025-05-22	Code
2	AdaFocus (Full, MViT-Charades-Pretrain-feature, MMN model)	89.4	No	Towards Weakly Supervised End-to-end Learning fo...	2023-11-28	-
3	MMN (Full, MViT-K400-Pretrain-feature, evaluated by AdaFocus)	88.3	No	Negative Sample Matters: A Renaissance of Metric...	2021-09-10	Code
4	AdaFocus (Full, I3D-Charades-Pretrain-feature, MMN model)	87.9	No	Towards Weakly Supervised End-to-end Learning fo...	2023-11-28	-
5	AdaFocus (Semi-weak, MViT-Charades-Pretrain-feature, D3G model)	86.1	No	Towards Weakly Supervised End-to-end Learning fo...	2023-11-28	-
6	MMN (Full, I3D-K400-Pretrain-feature, evaluated by AdaFocus)	85.8	No	Negative Sample Matters: A Renaissance of Metric...	2021-09-10	Code
7	AdaFocus (Weak, MViT-Charades-Pretrain-feature, CPL model)	85.2	No	Towards Weakly Supervised End-to-end Learning fo...	2023-11-28	-
8	CPL (Weak, MViT-K400-Pretrain-feature, evaluated by AdaFocus)	84.6	No	-	-	Code
9	AdaFocus (Weak, I3D-Charades-Pretrain-feature, CPL model)	84.2	No	Towards Weakly Supervised End-to-end Learning fo...	2023-11-28	-
10	D3G (Semi-weak, MViT-K400-Pretrain-feature, evaluated by AdaFocus)	83.1	No	D3G: Exploring Gaussian Prior for Temporal Sente...	2023-08-08	Code
11	CPL (Weak, I3D-K400-Pretrain-feature, evaluated by AdaFocus)	81.4	No	-	-	Code
12	AdaFocus (Semi-weak, I3D-Charades-Pretrain-feature, D3G model)	79.3	No	Towards Weakly Supervised End-to-end Learning fo...	2023-11-28	-
13	D3G (Semi-weak, I3D-K400-Pretrain-feature, evaluated by AdaFocus)	78.2	No	D3G: Exploring Gaussian Prior for Temporal Sente...	2023-08-08	Code

#1DeCafNetSOTA
91.53
R5@0.5· 2025-05-22
DeCafNet: Delegate and Conquer for Efficient Temporal Grounding in Long Videos Code
#2AdaFocus (Full, MViT-Charades-Pretrain-feature, MMN model)SOTA
89.4
R5@0.5· 2023-11-28
Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition
#3MMN (Full, MViT-K400-Pretrain-feature, evaluated by AdaFocus)SOTA
88.3
R5@0.5· 2021-09-10
Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding Code
#4AdaFocus (Full, I3D-Charades-Pretrain-feature, MMN model)
87.9
R5@0.5· 2023-11-28
Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition
#5AdaFocus (Semi-weak, MViT-Charades-Pretrain-feature, D3G model)
86.1
R5@0.5· 2023-11-28
Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition
#6MMN (Full, I3D-K400-Pretrain-feature, evaluated by AdaFocus)
85.8
R5@0.5· 2021-09-10
Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding Code
#7AdaFocus (Weak, MViT-Charades-Pretrain-feature, CPL model)
85.2
R5@0.5· 2023-11-28
Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition
#8CPL (Weak, MViT-K400-Pretrain-feature, evaluated by AdaFocus)
84.6
R5@0.5
No paperCode
#9AdaFocus (Weak, I3D-Charades-Pretrain-feature, CPL model)
84.2
R5@0.5· 2023-11-28
Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition
#10D3G (Semi-weak, MViT-K400-Pretrain-feature, evaluated by AdaFocus)
83.1
R5@0.5· 2023-08-08
D3G: Exploring Gaussian Prior for Temporal Sentence Grounding with Glance Annotation Code
#11CPL (Weak, I3D-K400-Pretrain-feature, evaluated by AdaFocus)
81.4
R5@0.5
No paperCode
#12AdaFocus (Semi-weak, I3D-Charades-Pretrain-feature, D3G model)
79.3
R5@0.5· 2023-11-28
Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition
#13D3G (Semi-weak, I3D-K400-Pretrain-feature, evaluated by AdaFocus)
78.2
R5@0.5· 2023-08-08
D3G: Exploring Gaussian Prior for Temporal Sentence Grounding with Glance Annotation Code