List your company

Temporal Action Detection on ActivityNet 1.3

64.2mAP@0.5MambaTAD
51.907255.098658.2961.4814Jun 18, 2021Apr 17, 2022Feb 14, 2023Dec 14, 2023Oct 12, 2024Aug 11, 2025Jun 10, 2026

Evaluation results

MethodLinks
MambaTADE2E=true, Mem=16.7G, Backbone=VideoMAEv2-G2025.1164.24511.343.8
ESTF-SSMBackbone=VideoMAEv2, E2E=true, Mem=28.6G2026.0463.844.611.243.9
MambaTADE2E=true, Mem=9.3G, Backbone=VideoMAE-H2025.1163.344.411.243.1
MambaTADE2E=false, Backbone=IV6B2025.1163.144.21142.8
MambaTADBackbone=InternVideo2, E2E=true, Mem=30.7G2026.0463.144.21143.5
ActionMambaTraining Type=Head only, Head Type=Anchor-free, Feature=InternVideo22025.0562.443.510.242
VideoMambaSuiteE2E=false, Backbone=IV6B2025.1162.443.510.242
ActionMambaBackbone=InternVideo2, E2E=false2026.0462.237.19.641.7
VideoMambaBackbone=InternVideo2, E2E=false2026.0462.143.29.541.6
DIGITTraining Type=Head only, Head Type=Query-based, Feature=InternVideo22025.056243.111.342
AdaTADTraining Type=Full, Feature=VideoMAEv2-g2025.0561.743.410.941.9
ActionFormerTraining Type=Head only, Head Type=Anchor-free, Feature=InternVideo22025.0561.544.612.741.2
InternVideo2*E2E=false, Backbone=IV6B2025.1161.544.612.741.2
TE-TADTraining Type=Head only, Head Type=Query-based, Feature=InternVideo22025.0561.341.810.941.1
AdaTAD++Backbone=VideoMAEv2, E2E=true, Mem=30.9G2026.0460.343.210.141
TadTRTraining Type=Head only, Head Type=Query-based, Feature=InternVideo2, MDGE=true, CAID=true2025.0560.24111.240.5
MambaTADE2E=false, Backbone=R(2+1)D2025.1160.241.39.740.2
AdaTADBackbone=VideoMAEv2, E2E=true, Mem=29.9G2026.0459.8429.740.8
BDRC-NetBackbone=InternVideo2, E2E=false2026.0459.638.99.841.9
LFAFE2E=false, Backbone=I3D†2025.1159.339.87.839
DIGITTraining Type=Head only, Head Type=Query-based, Feature=InternVideo2, NMS=false2025.0558.943.411.441.3
CLTDR-GMGBackbone=InternVideo2, E2E=false2026.0458.340.29.439.5
DyFADetBackbone=TSP+IV12026.0458.1939.38.6338.62
Zhu et al.Backbone=I3D2024.0358.136.36.235.2
DyFADetTraining Type=Head only, Head Type=Anchor-free, Feature=I3D / R(2+1)D2025.0558.139.68.438.5
DyFADetE2E=false, Backbone=R(2+1)D2025.1158.139.68.438.5
AdaTADE2E=true, Mem=19.2G, Backbone=VideoMAE-H2025.115840.69.839.4
AdaTADE2E=true, Mem=11.0G, Backbone=VideoMAE-L2025.1157.740.610.139.2
AdaTADType=ANN, Backbone=ViT-L [17]2026.0657.6940.5610.1339.22
TadTRTraining Type=Head only, Head Type=Query-based, Feature=InternVideo22025.0557.138.81138.2
AdaTADE2E=true, Mem=4.9G, Backbone=VideoMAE-B2025.1156.839.49.738.4
AdaTADType=ANN, Backbone=ViT-B [17]2026.0656.7739.359.7138.39
TriDetBackbone=SlowFast, Flow=false2023.1256.739.311.738.6
TAGSBackbone=I3D2024.0356.336.89.636.5
AdaTADE2E=true, Mem=2.5G, Backbone=VideoMAE-S2025.1156.2399.137.9
AdaTADType=ANN, Backbone=ViT-S [17]2026.0656.1538.999.0737.85
DiffTADE2E=false, Backbone=I3D†2025.1156.136.9936.1
DiffTAD (10-step)Backbone=I3D2026.0456.136.9936.1
Baseline(ANN)Type=ANN, Backbone=ViT-S [17]2026.0656.0938.718.8637.64
SpikeTADType=ANN2SNN, Backbone=ViT-S [17], T (time-steps in the backbone)=162026.0656.0438.978.7537.42
ContextLocBackbone=I3D2024.035635.23.634.2
ViT-TADE2E=true, Backbone=ViT-B2025.1155.938.58.837.4
ViT-TADBackbone=ViT-B, Flow=false, Resolution=160x1602023.1255.8738.478.837.4
ViT-TADType=ANN, Backbone=ViT-B [17]2026.0655.8738.478.837.4
SpikeTADType=ANN2SNN, Backbone=ViT-S [17], T (time-steps in the backbone)=82026.0655.8737.998.6937.05
SpikeTAD(ANN)Type=ANN, Backbone=ViT-S [17]2026.0655.8337.928.6737.12
ViT-TADTraining Type=Full, Feature=ViT-B2025.0555.838.58.837.4
CausalTADBackbone=TSP2026.0455.6238.519.437.46
ADSFormerE2E=false, Backbone=VM22025.1155.338.48.437.1
Re2TALType=ANN, Backbone=Re2SlowFast-101 [13]2026.0655.337.99.137
SpikeTADType=ANN2SNN, Backbone=ViT-S [17], T (time-steps in the backbone)=42026.0655.2637.218.836.53
Re2TALBackbone=SlowFast-R101, Flow=false, Resolution=224x2242023.1255.2537.869.0537.01
LiquidTADBackbone=TSP2026.0455.1838.18.4337.06
ViT-TADBackbone=ViT-S, Flow=false, Resolution=160x1602023.1255.0937.818.7536.69
ViT-TADType=ANN, Backbone=ViT-S [17]2026.0655.0937.818.7536.69
ActionFormerBackbone=TSP2026.0455.0838.278.9137.07
TriDetBackbone=TSP2026.0454.8938.28.2136.96
Re2TALE2E=true, Mem=24G, Backbone=Re2VideoSwin-T2025.1154.837.8936.8
BSFTALBackbone=InternVideo2, E2E=false2026.0454.837.59.137.1
VSGNBackbone=TSP2026.0454.837.359.836.89
Re2TALBackbone=Swin-T, Flow=false, Resolution=224x2242023.1254.7537.819.0336.8
Re2TALType=ANN, Backbone=Re2VideoSwin-T [13]2026.0654.7537.819.0336.8
TriDetTraining Type=Head only, Head Type=Anchor-free, Feature=I3D / R(2+1)D2025.0554.7388.436.8
ActionFormerTraining Type=Head only, Head Type=Anchor-free, Feature=I3D / R(2+1)D2025.0554.737.88.436.6
ActionFormerE2E=false, Backbone=R(2+1)D2025.1154.737.88.436.6
TriDetE2E=false, Backbone=SlowFast2025.1154.7388.436.8
TridetType=ANN, Backbone=I3D [15]2026.0654.7388.436.8
TriDetBackbone=I3D, Flow=true2023.1254.536.811.536.8
DIGITTraining Type=Head only, Head Type=Query-based, Feature=I3D / R(2+1)D2025.0554.438.210.737.3
VSGN (w/o GCN)Backbone=TSP2026.0454.436.889.1536.27
DIGITTraining Type=Head only, Head Type=Query-based, Feature=I3D / R(2+1)D, NMS=false2025.0554.338.410.637.2
TCANetE2E=false, Backbone=SlowFast2025.1154.339.18.437.6
ActionFormerType=ANN, Backbone=SlowFast-R50 [43]2026.0654.2637.048.1336.02
TE-TADTraining Type=Head only, Head Type=Query-based, Feature=I3D / R(2+1)D2025.0554.238.110.637.1
TALLFormerBackbone=Swin-B, Flow=false, Resolution=224x2242023.1254.136.27.935.6
TALLFormerTraining Type=Full, Feature=Swin-B2025.0554.136.27.935.6
TALLFormerE2E=true, Mem=29G, Backbone=Swin2025.1154.136.27.935.6
TALLFormerType=ANN, Backbone=VideoSwin-B [45]2026.0654.136.27.935.6
PBRNetBackbone=I3D, Flow=true, Resolution=96x962023.1253.9634.978.9835.01
TadTR*Feature=TSP, Ensemble=true, focal loss=true, GFLOPs=0.0382021.0653.6237.5210.5636.75
TadTRBackbone=TSP2026.0453.6237.5210.5636.75
TadTRTraining Type=Head only, Head Type=Query-based, Feature=I3D / R(2+1)D2025.0553.637.510.636.8
TadTRE2E=false, Backbone=TSN†2025.1153.637.510.636.8
ActionFormerBackbone=I3D2024.0353.536.28.235.6
ActionFormerBackbone=I3D, Flow=true2023.1253.536.28.235.6
ActionFormerType=ANN, Backbone=I3D [15]2026.0653.536.28.235.6
BMNBackbone=TSP2026.0452.937.39.6736.4
TadTR*Feature=I3D, Ensemble=true, focal loss=true, GFLOPs=0.0382021.0652.8337.0510.8336.11
DualDETRBackbone=I3D2024.0352.6357.834.3
DualDETRTraining Type=Head only, Head Type=Query-based, Feature=I3D2025.0552.6357.834.3
TSIBackbone=TSP2026.0452.4435.579.835.36
BMN-CSABackbone=TSN2024.0352.436.25.235.4
VSGNBackbone=TSN2024.0352.4368.435.1
AFSDBackbone=I3D2024.0352.435.36.534.4
AFSDBackbone=I3D, Flow=true, Resolution=96x962023.1252.435.36.534.4
AFSDTraining Type=Full, Feature=I3D2025.0552.435.36.534.4
VSGNE2E=false, Backbone=TSN†2025.1152.4368.435.1
AFSDE2E=true, Mem=12G, Backbone=I3D†2025.1152.435.36.534.4
AFSDFeature=I3D, Ensemble=true, GFLOPs=3.32021.0652.3835.276.4734.39
VSGNBackbone=TSN, Flow=true2023.1252.3836.018.3735.07