Zero-Shot Learning on THUMOS’14

Metric: Avg mAP (0.3:0.7) (higher is better)

LeaderboardDataset

Loading chart...

Results

Hide augmentations

Sort:

#	Model↕	Avg mAP (0.3:0.7)▼	Augmentations	Paper	Date↕	Code
1	AdaTAD (VideoMAEv2-giant)	76.9	No	End-to-End Temporal Action Detection with 1B Par...	2023-11-28	Code
2	RDFA-S6 (InternVideo2-6B)	74.2	No	Enhancing Temporal Action Localization: Advanced...	2024-07-18	Code
3	ActionMamba(InternVideo2-6B)	72.72	No	Video Mamba Suite: State Space Model as a Versat...	2024-03-14	Code
4	InternVideo2-6B	72	Yes	InternVideo2: Scaling Foundation Models for Mult...	2024-03-22	Code
5	ActionFormer (InternVideo features)	71.58	Yes	InternVideo: General Video Foundation Models via...	2022-12-06	Code
6	TriDet (VideoMAE v2-g feature)	70.1	Yes	Temporal Action Localization with Enhanced Insta...	2023-09-11	Code
7	InternVideo2-1B	69.8	Yes	InternVideo2: Scaling Foundation Models for Mult...	2024-03-22	Code
8	ActionFormer (VideoMAE V2-g features)	69.6	Yes	VideoMAE V2: Scaling Video Masked Autoencoders w...	2023-03-29	Code
9	TriDet (I3D features)	69.3	No	TriDet: Temporal Action Detection with Relative ...	2023-03-13	Code
10	ASL(I3D features)	67.9	No	Action Sensitivity Learning for Temporal Action ...	2023-05-25	-
11	TemporalMaxer (I3D features)	67.7	No	TemporalMaxer: Maximize Temporal Context with on...	2023-03-16	Code
12	DualDETR (I3D features)	66.8	No	Dual DETRs for Multi-Label Temporal Action Detec...	2024-03-31	-
13	ActionFormer (I3D features)	66.8	No	ActionFormer: Localizing Moments of Actions with...	2022-02-16	Code
14	TadML(two-stream)	59.7	No	TadML: A fast temporal action detection with Mec...	2022-06-07	Code
15	BasicTAD (160,6,192,R50-SlowOnly)	59.6	No	BasicTAD: an Astounding RGB-Only Baseline for Te...	2022-05-05	Code
16	TadTR	56.7	No	End-to-end Temporal Action Detection with Transf...	2021-06-18	Code
17	ReAct (TSN features)	55	No	ReAct: Temporal Action Detection with Relational...	2022-07-14	Code
18	BasicTAD (112,3,96,R50-SlowOnly)	54.9	No	BasicTAD: an Astounding RGB-Only Baseline for Te...	2022-05-05	Code
19	E2E-TAD (SlowFast R50+TadTR)	54.2	No	An Empirical Study of End-to-End Temporal Action...	2022-04-06	Code
20	TadML(rgb-only)	53.46	No	TadML: A fast temporal action detection with Mec...	2022-06-07	Code
21	MUSES	53.4	No	Multi-shot Temporal Event Localization: a Benchm...	2020-12-17	Code
22	AVFusion	53.3	Yes	Hear Me Out: Fusional Approaches for Audio Augme...	2021-06-27	Code
23	TAGS (I3D)	52.8	No	Proposal-Free Temporal Action Detection via Glob...	2022-07-14	Code
24	DCAN (TSN features)	52.3	No	DCAN: Improving Temporal Action Detection via Du...	2021-12-07	Code
25	TSP	50.46	No	TSP: Temporally-Sensitive Pretraining of Video E...	2020-11-23	Code
26	VSGN	50.2	No	Video Self-Stitching Graph Network for Temporal ...	2020-11-30	Code
27	DaoTAD	50	No	RGB Stream Is Enough for Temporal Action Detection	2021-07-09	Code
28	Decouple-SSAD	42	No	Decoupling Localization and Classification in Si...	2019-04-16	Code
29	TAL-Net	39.8	No	Rethinking the Faster R-CNN Architecture for Tem...	2018-04-20	-

#1AdaTAD (VideoMAEv2-giant)SOTA
76.9
Avg mAP (0.3:0.7)· 2023-11-28
End-to-End Temporal Action Detection with 1B Parameters Across 1000 Frames Code
#2RDFA-S6 (InternVideo2-6B)
74.2
Avg mAP (0.3:0.7)· 2024-07-18
Enhancing Temporal Action Localization: Advanced S6 Modeling with Recurrent Mechanism Code
#3ActionMamba(InternVideo2-6B)
72.72
Avg mAP (0.3:0.7)· 2024-03-14
Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding Code
#4InternVideo2-6B
72
Avg mAP (0.3:0.7)· Augmentations· 2024-03-22
InternVideo2: Scaling Foundation Models for Multimodal Video Understanding Code
#5ActionFormer (InternVideo features)SOTA
71.58
Avg mAP (0.3:0.7)· Augmentations· 2022-12-06
InternVideo: General Video Foundation Models via Generative and Discriminative Learning Code
#6TriDet (VideoMAE v2-g feature)
70.1
Avg mAP (0.3:0.7)· Augmentations· 2023-09-11
Temporal Action Localization with Enhanced Instant Discriminability Code
#7InternVideo2-1B
69.8
Avg mAP (0.3:0.7)· Augmentations· 2024-03-22
InternVideo2: Scaling Foundation Models for Multimodal Video Understanding Code
#8ActionFormer (VideoMAE V2-g features)
69.6
Avg mAP (0.3:0.7)· Augmentations· 2023-03-29
VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking Code
#9TriDet (I3D features)
69.3
Avg mAP (0.3:0.7)· 2023-03-13
TriDet: Temporal Action Detection with Relative Boundary Modeling Code
#10ASL(I3D features)
67.9
Avg mAP (0.3:0.7)· 2023-05-25
Action Sensitivity Learning for Temporal Action Localization
#11TemporalMaxer (I3D features)
67.7
Avg mAP (0.3:0.7)· 2023-03-16
TemporalMaxer: Maximize Temporal Context with only Max Pooling for Temporal Action Localization Code
#12DualDETR (I3D features)
66.8
Avg mAP (0.3:0.7)· 2024-03-31
Dual DETRs for Multi-Label Temporal Action Detection
#13ActionFormer (I3D features)SOTA
66.8
Avg mAP (0.3:0.7)· 2022-02-16
ActionFormer: Localizing Moments of Actions with Transformers Code
#14TadML(two-stream)
59.7
Avg mAP (0.3:0.7)· 2022-06-07
TadML: A fast temporal action detection with Mechanics-MLP Code
#15BasicTAD (160,6,192,R50-SlowOnly)
59.6
Avg mAP (0.3:0.7)· 2022-05-05
BasicTAD: an Astounding RGB-Only Baseline for Temporal Action Detection Code
#16TadTRSOTA
56.7
Avg mAP (0.3:0.7)· 2021-06-18
End-to-end Temporal Action Detection with Transformer Code
#17ReAct (TSN features)
55
Avg mAP (0.3:0.7)· 2022-07-14
ReAct: Temporal Action Detection with Relational Queries Code
#18BasicTAD (112,3,96,R50-SlowOnly)
54.9
Avg mAP (0.3:0.7)· 2022-05-05
BasicTAD: an Astounding RGB-Only Baseline for Temporal Action Detection Code
#19E2E-TAD (SlowFast R50+TadTR)
54.2
Avg mAP (0.3:0.7)· 2022-04-06
An Empirical Study of End-to-End Temporal Action Detection Code
#20TadML(rgb-only)
53.46
Avg mAP (0.3:0.7)· 2022-06-07
TadML: A fast temporal action detection with Mechanics-MLP Code
#21MUSESSOTA
53.4
Avg mAP (0.3:0.7)· 2020-12-17
Multi-shot Temporal Event Localization: a Benchmark Code
#22AVFusion
53.3
Avg mAP (0.3:0.7)· Augmentations· 2021-06-27
Hear Me Out: Fusional Approaches for Audio Augmented Temporal Action Localization Code
#23TAGS (I3D)
52.8
Avg mAP (0.3:0.7)· 2022-07-14
Proposal-Free Temporal Action Detection via Global Segmentation Mask Learning Code
#24DCAN (TSN features)
52.3
Avg mAP (0.3:0.7)· 2021-12-07
DCAN: Improving Temporal Action Detection via Dual Context Aggregation Code
#25TSPSOTA
50.46
Avg mAP (0.3:0.7)· 2020-11-23
TSP: Temporally-Sensitive Pretraining of Video Encoders for Localization Tasks Code
#26VSGN
50.2
Avg mAP (0.3:0.7)· 2020-11-30
Video Self-Stitching Graph Network for Temporal Action Localization Code
#27DaoTAD
50
Avg mAP (0.3:0.7)· 2021-07-09
RGB Stream Is Enough for Temporal Action Detection Code
#28Decouple-SSADSOTA
42
Avg mAP (0.3:0.7)· 2019-04-16
Decoupling Localization and Classification in Single Shot Temporal Action Detection Code
#29TAL-NetSOTA
39.8
Avg mAP (0.3:0.7)· 2018-04-20
Rethinking the Faster R-CNN Architecture for Temporal Action Localization