List your company
ResearchFormerly WizWand

Video Action Classification on Kinetics-400

0.894Top-1 AccuracyInternVideo2_s2-1B
0.357360.496680.6360.77532Aug 9, 2020May 30, 2021Mar 21, 2022Jan 9, 2023Oct 31, 2023Aug 20, 2024Jun 11, 2025
Updated 8mo ago

Evaluation results

MethodLinks
InternVideo2_s2-1BParameters=1B, Evaluation Protocol=Attentive probe2025.060.894—
InternVideo2-6BParameters=6B2025.060.888—
PEcoreGParameters=1.9B, Evaluation Protocol=Attentive probe2025.060.885—
InternVideo2-1BParameters=1B2025.060.879—
VideoPrismParameters=1B2025.060.876—
SigLIP2Parameters=1.2B, Evaluation Protocol=Attentive probe2025.060.873—
V-JEPA 2 ViT-g384Parameters=1B, Input Resolution=384x384, Input format=16x8x3 frames2025.060.873—
Video-SwinV2-Gtrain I(W) size=320(20)^2 x 8(8), test I(W) size=384(24)^2 x 8(8), views=4x52021.110.868—
V-JEPA 2 ViT-gParameters=1B, Input format=16x8x3 frames2025.060.866—
TokenLearnertrain I(W) size=256(8)^2 x 64(64), test I(W) size=256(8)^2 x 64(64), views=4x32021.110.854—
V-JEPA 2 ViT-HParameters=600M, Input format=16x8x3 frames2025.060.853—
V-JEPA 2 ViT-LParameters=300M, Input format=16x8x3 frames2025.060.851—
SwinV1-Ltrain I(W) size=480(12)^2 x 16(8), test I(W) size=480(12)^2 x 16(8), views=10x52021.110.849—
ViViTviews=4x32021.110.848—
V-JEPA ViT-HParameters=600M, Evaluation Protocol=Attentive probe2025.060.845—
DINOv2Parameters=1.1B, Evaluation Protocol=Attentive probe2025.060.836—
Video-SwinV2-Gtrain I(W) size=320(20)^2 x 8(8), test I(W) size=384(24)^2 x 8(8), views=1x12021.110.834—
Video-SwinV2-Gtrain I(W) size=320(20)^2 x 8(8), test I(W) size=320(20)^2 x 8(8), views=1x12021.110.832—
VideoMAEv2Parameters=1B2025.060.828—
VATT + BothPre-training Dataset=HT100M + AudioSet + YT8M, Cross-Modality Gradient Realignment (GR)=true, Gradient-based Curriculum Learning (CL)=true2022.110.80010.9469
VATT + CLPre-training Dataset=HT100M + AudioSet, Gradient-based Curriculum Learning (CL)=true2022.110.79890.9471
VATT + GRPre-training Dataset=HT100M + AudioSet + YT8M, Cross-Modality Gradient Realignment (GR)=true2022.110.79730.9457
VATT + CLPre-training Dataset=HT100M + AudioSet + YT8M, Gradient-based Curriculum Learning (CL)=true2022.110.7970.948
VATTPre-training Dataset=HT100M + AudioSet + YT8M2022.110.79390.9456
VATT + GRPre-training Dataset=HT100M + AudioSet, Cross-Modality Gradient Realignment (GR)=true2022.110.79290.9432
VATT + BothPre-training Dataset=HT100M + AudioSet, Cross-Modality Gradient Realignment (GR)=true, Gradient-based Curriculum Learning (CL)=true2022.110.79260.9448
VATTPre-training Dataset=HT100M + AudioSet2022.110.79230.943
VATT + RW (VT)Pre-training Dataset=HT100M + AudioSet, Gradient Re-weighting (RW)=Video-Text2022.110.78590.9417
VATT + RW (VT)Pre-training Dataset=HT100M + AudioSet + YT8M, Gradient Re-weighting (RW)=Video-Text2022.110.78430.9438
VATT + RW (VA)Pre-training Dataset=HT100M + AudioSet + YT8M, Gradient Re-weighting (RW)=Video-Audio2022.110.7770.9383
VATT + CLPre-training Dataset=HT100M, Gradient-based Curriculum Learning (CL)=true2022.110.77250.9338
VATT + GRPre-training Dataset=HT100M, Cross-Modality Gradient Realignment (GR)=true2022.110.76720.9272
VATT + BothPre-training Dataset=HT100M, Cross-Modality Gradient Realignment (GR)=true, Gradient-based Curriculum Learning (CL)=true2022.110.76590.9326
VATT + RW (VA)Pre-training Dataset=HT100M + AudioSet, Gradient Re-weighting (RW)=Video-Audio2022.110.76560.9352
Supervised (K400)Backbone=R3D-50 (31.7M), Pre-train data=N/A, Modality=V, Evaluation Protocol=N/A2020.080.76—
VATT + RW (VT)Pre-training Dataset=HT100M, Gradient Re-weighting (RW)=Video-Text2022.110.75660.928
VATTPre-training Dataset=HT100M2022.110.74710.9269
CVRLBackbone=R3D-152 (2x) (328.0M), Pre-train data=K600 (11d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.716—
VATT + RW (VA)Pre-training Dataset=HT100M, Gradient Re-weighting (RW)=Video-Audio2022.110.70550.9041
CVRLBackbone=R3D-101 (59.7M), Pre-train data=K400 (28d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.676—
CVRLBackbone=R3D-50 (31.7M), Pre-train data=K400 (28d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.661—
SeCoBackbone=R-50 (23.5M), Pre-train data=K400 (28d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.619—
TANSettings=finetuned with TAN, Backbone=S3D, Evaluation Protocol=Linear Probing2022.040.562—
MIL-NCESettings=reproduce of [47], Backbone=S3D, Evaluation Protocol=Linear Probing2022.040.557—
ImageNet inflatedBackbone=R3D-50 (31.7M), Pre-train data=ImageNet (N/A), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.535—
VINCEBackbone=R-50 (23.5M), Pre-train data=K400 (28d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.491—
SimCLR inflatedBackbone=R3D-50 (31.7M), Pre-train data=K400 (28d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.468—
VTHCLBackbone=R3D-50 (31.7M), Pre-train data=K400 (28d), Modality=V, Evaluation Protocol=Linear eval on K4002020.080.378—