List your company
ResearchFormerly WizWand

Video Action Recognition on HMDB51 (Top-1 Accuracy)

88.1Top-1 AccuracyVideoMAE V2
29.96445.05760.1575.243Jun 18, 2021Mar 27, 2022Jan 4, 2023Oct 14, 2023Jul 23, 2024May 2, 2025Feb 9, 2026
Updated 5mo ago

Evaluation results

MethodLinks
VideoMAE V22025.0988.1
OV-Encoder (Codec)Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0285.3
OV-Encoder (Codec)Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0283.4
OV-Encoder (Frame)Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0283.1
MoTIFBackbone=PE-G/142025.0983
AIMv2Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0282.6
MoTIF-STBackbone=PE-G/142025.0982.1
MoTIFBackbone=PE-L/142025.0981.8
OV-Encoder (Frame)Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0281.6
AIMv2Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0281.3
SigLIP2Backbone=ViT-L/16, Resolution=256, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0281.2
DINOv3Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0279.7
SigLIP2Backbone=ViT-L/16, Resolution=256, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0279.1
SigLIPBackbone=ViT-L/16, Resolution=256, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0278.8
DINOv3Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0278.6
MetaCLIP2Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=16 Frames / 4096 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0278.2
CLIPBackbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0278
Global CBMBackbone=PE-G/14, Evaluation Protocol=Supervised2025.0977.8
MetaCLIPBackbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0277.1
SigLIPBackbone=ViT-L/16, Resolution=256, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0277
MetaCLIP2Backbone=ViT-L/14, Resolution=224, Input Configuration (Frames/Patches)=8 Frames / 2048 Patches, Evaluation Protocol (Attentive probe with frozen backbones)=Attentive probe with frozen backbones2026.0276.3
MoTIFBackbone=ViT-L/142025.0976.1
TSM2025.0973.5
NoFrame2025.0973.4
OmniVLEvaluation Protocol=linear probing2022.0970
FiTEvaluation Protocol=linear probing2022.0968.8
ELoEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=S3D, Param=8.8M, Res=224, Frames=32, Modality=V+T2021.0667.4
GDTEvaluation Protocol=Finetune, Pre-train Dataset=Audioset, Backbone=R(2+1)D, Param=14.4M, Res=224, Frames=32, Modality=V+A2021.0666.1
GDTEvaluation Protocol=Finetune, Pretrain Dataset=Audioset, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=224, Frames=32, Modality=V+A2021.0666.1
CVRLEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R3D-50, Param=36.1M, Res=224, Frames=16, Modality=V2021.0665.4
CoCLR †Evaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0662.9
CoCLR †Evaluation Protocol=Finetune, Pretrain Dataset=K-400, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0662.9
ViCC-R+FEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0662.2
ViCC-R+FEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0662.2
ViCC-R+FEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=128, Frames=16, Modality=V2021.0661.5
ViCC-R+FEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=128, Frames=16, Modality=V2021.0661.5
MotionFitEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=32, Modality=V2021.0661.4
MIL-NCEEvaluation Protocol=Finetune, Pre-train Dataset=HTM, Backbone=S3D, Param=8.8M, Res=224, Frames=32, Modality=V+T2021.0661
MIL-NCEEvaluation Protocol=Finetune, Pretrain Dataset=HTM, Backbone=S3D, Number of Parameters=8.8M, Resolution=224, Frames=32, Modality=V+T2021.0661
Zero-shotBackbone=PE-G/14, Evaluation Protocol=Zero-shot2025.0960.7
TimeSformerPre-training=Sup21K, Evaluation Protocol=linear probing2022.0960.1
CoCLR †Evaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0658.7
CoCLR †Evaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0658.7
AdaptFormer-64Avg. Params (M)=1.262022.0555.69
ViCC-R+FEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R3D, Param=14.2M, Res=128, Frames=16, Modality=V2021.0653.2
SeLaViEvaluation Protocol=Finetune, Pre-train Dataset=VGG-sound, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=30, Modality=V+A2021.0653.1
MIL-NCEEvaluation Protocol=Linear, Pre-train Dataset=HTM, Backbone=S3D, Param=8.8M, Res=224, Frames=32, Modality=V+T2021.0653.1
SeLaViEvaluation Protocol=Finetune, Pretrain Dataset=VGG-sound, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=112, Frames=30, Modality=V+A2021.0653.1
MIL-NCEEvaluation Protocol=Linear, Pretrain Dataset=HTM, Backbone=S3D, Number of Parameters=8.8M, Resolution=224, Frames=32, Modality=V+T2021.0653.1
VPTAvg. Params (M)=0.082022.0552.67
XDCEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R(2+1)D, Param=14.4M, Res=224, Frames=32, Modality=V+A2021.0652.6
Pace PredEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=S3D-G, Param=9.6M, Res=224, Frames=64, Modality=V2021.0652.6
XDCEvaluation Protocol=Finetune, Pretrain Dataset=K-400, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=224, Frames=32, Modality=V+A2021.0652.6
Pace PredEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=S3D-G, Number of Parameters=9.6M, Resolution=224, Frames=64, Modality=V2021.0652.6
ViCC-RGBEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=128, Frames=16, Modality=V2021.0652.4
CoCLR †Evaluation Protocol=Linear, Pre-train Dataset=K-400, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0652.4
ViCC-RGBEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=128, Frames=16, Modality=V2021.0652.4
CoCLR †Evaluation Protocol=Linear, Pretrain Dataset=K-400, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0652.4
CoCLREvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0652.1
CoCLREvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0652.1
AdaptFormer-4Avg. Params (M)=0.152022.0551.81
AdaptFormer-1Avg. Params (M)=0.102022.0551.68
LinearAvg. Params (M)=0.072022.0549.84
SpeedNetEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=S3D-G, Param=8.8M, Res=128, Frames=32, Modality=V2021.0648.8
SpeedNetEvaluation Protocol=Finetune, Pretrain Dataset=K-400, Backbone=S3D-G, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0648.8
ViCC-RGBEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0647.9
ViCC-R+FEvaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0647.9
ViCC-RGBEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0647.9
ViCC-R+FEvaluation Protocol=Linear, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0647.9
RTTEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R3D, Param=14.2M, Res=112, Frames=16, Modality=V2021.0647.5
ViCC-R+FEvaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=R3D, Param=14.2M, Res=128, Frames=16, Modality=V2021.0646.7
Full-tuningAvg. Params (M)=86.042022.0546.41
RTTEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=16, Modality=V2021.0646.4
RTTEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=112, Frames=16, Modality=V2021.0646.4
ViCC-R+FEvaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=128, Frames=16, Modality=V2021.0645.2
ViCC-R+FEvaluation Protocol=Linear, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=128, Frames=16, Modality=V2021.0645.2
VIEEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R3D, Param=14.2M, Res=112, Frames=40, Modality=V2021.0644.8
ViCC-RGBEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R3D, Param=14.2M, Res=128, Frames=16, Modality=V2021.0644.7
CBTEvaluation Protocol=Finetune, Pre-train Dataset=K-600, Backbone=S3D, Param=8.8M, Res=112, Frames=16, Modality=V+T2021.0644.6
CBTEvaluation Protocol=Finetune, Pretrain Dataset=K-600, Backbone=S3D, Number of Parameters=8.8M, Resolution=112, Frames=16, Modality=V+T2021.0644.6
ViCC-RGBEvaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=R3D, Param=14.2M, Res=128, Frames=16, Modality=V2021.0644.2
CoCLR †Evaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0640.2
CoCLR †Evaluation Protocol=Linear, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0640.2
CoCLREvaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0639.1
CoCLREvaluation Protocol=Linear, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0639.1
ViCC-RGBEvaluation Protocol=Linear, Pre-train Dataset=UCF101, Backbone=S3D, Param=8.8M, Res=128, Frames=32, Modality=V2021.0638.5
ViCC-RGBEvaluation Protocol=Linear, Pretrain Dataset=UCF101, Backbone=S3D, Number of Parameters=8.8M, Resolution=128, Frames=32, Modality=V2021.0638.5
Var. PSPEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=16, Modality=V2021.0636.8
Var. PSPEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=112, Frames=16, Modality=V2021.0636.8
Pace PredEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=16, Modality=V2021.0636.6
Pace PredEvaluation Protocol=Finetune, Pretrain Dataset=K-400, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=112, Frames=16, Modality=V2021.0636.6
Pace PredEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=16, Modality=V2021.0635.9
Pace PredEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=112, Frames=16, Modality=V2021.0635.9
PRPEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=16, Modality=V2021.0635
PRPEvaluation Protocol=Finetune, Pretrain Dataset=UCF101, Backbone=R(2+1)D, Number of Parameters=14.4M, Resolution=112, Frames=16, Modality=V2021.0635
DPCEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R3D, Param=14.2M, Res=128, Frames=40, Modality=V2021.0634.5
Var. PSPEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R3D, Param=14.2M, Res=112, Frames=16, Modality=V2021.0633.7
RotNet3DEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R3D, Param=33.6M, Res=224, Frames=16, Modality=V2021.0633.7
ST-PuzzleEvaluation Protocol=Finetune, Pre-train Dataset=K-400, Backbone=R3D, Param=33.6M, Res=224, Frames=16, Modality=V2021.0633.7
VCPEvaluation Protocol=Finetune, Pre-train Dataset=UCF101, Backbone=R(2+1)D, Param=14.4M, Res=112, Frames=16, Modality=V2021.0632.2