Home News Latest Papers Tutorials Datasets Wiki SOTA LLM Models GPU Leaderboard Events

English

Temporal Action Localization On Hacs

Metrics

Average-mAP

mAP@0.5

mAP@0.75

mAP@0.95

Results

Performance results of various models on this benchmark

Model Name	Average-mAP	mAP@0.5	mAP@0.75	mAP@0.95	Paper Title	Repository
TadTr (I3D RGB)	32.09	47.14	32.11	10.94	End-to-end Temporal Action Detection with Transformer
TriDet (VideoMAEv2)	43.1	62.4	44.1	13.1	Temporal Action Localization with Enhanced Instant Discriminability
DyFADet(VideoMAEv2)	44.3	64.0	44.8	14.1	DyFADet: Dynamic Feature Aggregation for Temporal Action Detection
ActionMamba(InternVideo2-6B)	44.56	64.02	45.71	13.34	Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding
InternVideo2-1B	42.4	-	-	-	InternVideo2: Scaling Foundation Models for Multimodal Video Understanding
TriDet (SlowFast)	38.6	56.7	39.3	11.7	TriDet: Temporal Action Detection with Relative Boundary Modeling
InternVideo2-6B	43.3	-	-	-	InternVideo2: Scaling Foundation Models for Multimodal Video Understanding
SSN	18.97	28.82	18.80	5.32	HACS: Human Action Clips and Segments Dataset for Recognition and Temporal Localization
TriDet (I3D RGB)	36.8	54.5	36.8	11.5	TriDet: Temporal Action Detection with Relative Boundary Modeling
RDFA-S6 (InternVideo2-6B)	45.8	66.4	47.2	14.3	Enhancing Temporal Action Localization: Advanced S6 Modeling with Recurrent Mechanism	-
LoFi+G-TAD (RGB, RN18)	24.64	37.78	24.40	7.29	Low-Fidelity Video Encoder Optimization for Temporal Action Localization	-
InternVideo	41.55	-	-	-	InternVideo: General Video Foundation Models via Generative and Discriminative Learning

0 of 12 row(s) selected.