List your company
ResearchFormerly WizWand

Action Recognition on K400

96.7Top-1 AccuracyVideo-STAR-7B
70.1877.06583.9590.835Oct 9, 2025Nov 6, 2025Dec 5, 2025Jan 3, 2026Feb 1, 2026Mar 2, 2026Mar 31, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Video-STAR-7BProtocol=Base-to-Novel (harmonic mean)2026.0396.7
Video-STAR-7B2025.1096.7
Gemini-1.5-Pro2025.1092.8
InternVideo2Protocol=Fine-Tuned2026.0392.1
Qwen3-VL-8B2025.1089.5
InternVideo2s2-1BParam.=1B, Resolution=256x256, Frame clips=16, Temporal crops=8, Spatial crops=32026.0389.4
PEcore GParam.=1.9B, Resolution=256x256, Frame clips=16, Temporal crops=8, Spatial crops=32026.0388.5
InternVideo2Type=Discriminative Pretraining, Pretrained on=Videos, Arch.=ViT-G, Param.=1B, Attentive probing=true2025.1287.9
InternVideo2-1BParam.=1B2026.0387.9
DINOv3Param.=7B2026.0387.8
V-JEPA 2.1 ViT-GParam.=2B, Resolution=384x384, Frame clips=16, Temporal crops=8, Spatial crops=32026.0387.7
VideoPrismType=Discriminative Pretraining, Pretrained on=Videos, Arch.=ViT-G, Param.=1B, Attentive probing=true2025.1287.6
VideoPrismParam.=1B2026.0387.6
Siglip2Type=Discriminative Pretraining, Pretrained on=Images, Arch.=ViT-G, Param.=1.2B, Attentive probing=true2025.1287.3
SigLIP2Param.=1.2B, Resolution=256x256, Frame clips=16, Temporal crops=8, Spatial crops=32026.0387.3
V-JEPA 2 ViT-gParam.=1B, Resolution=384x384, Frame clips=16, Temporal crops=8, Spatial crops=32026.0387.3
V-JEPA 2.1 ViT-gParam.=1B, Resolution=384x384, Frame clips=16, Temporal crops=8, Spatial crops=32026.0387
DINOv3 ViT-H+Param.=0.8B2026.0386.7
VJEPA2Type=Discriminative Pretraining, Pretrained on=Videos, Arch.=ViT-G, Param.=1B, Attentive probing=true2025.1286.6
Qwen2.5-VL-7B2025.1086.3
OmniStream2026.0385.7
V-JEPA2-LModel Size=Large2026.0385.1
V-JEPA ViT-HParam.=600M, Resolution=256x256, Frame clips=16, Temporal crops=8, Spatial crops=32026.0384.5
DINOv3-LModel Size=Large2026.0383.6
DINOv2Param.=1.1B, Resolution=256x256, Frame clips=16, Temporal crops=8, Spatial crops=32026.0383.6
DINOv2Type=Discriminative Pretraining, Pretrained on=Images, Arch.=ViT-G, Param.=1.1B, Attentive probing=true2025.1283.4
NExT-Vid-GType=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-G, Param.=1.1B, Attentive probing=true2025.1283.1
VideoMAEv2Param.=1B2026.0382.8
VJEPAType=Discriminative Pretraining, Pretrained on=Videos, Arch.=ViT-H, Param.=630M, Attentive probing=true2025.1282
OpenCLIPType=Discriminative Pretraining, Pretrained on=Images, Arch.=ViT-G, Param.=1.8B, Attentive probing=true2025.1281.8
NExT-Vid-HType=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-H, Param.=600M, Attentive probing=true2025.1280.6
VideoMAEType=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-H, Param.=630M, Attentive probing=true2025.1279.8
IJEPAType=Discriminative Pretraining, Pretrained on=Images, Arch.=ViT-H, Param.=630M, Attentive probing=true2025.1279.7
MVDType=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-L, Param.=200M, Attentive probing=true2025.1279.4
NExT-Vid-LType=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-L, Param.=300M, Attentive probing=true2025.1278.5
TotoType=Generative Pretraining, Pretrained on=Videos, Arch.=LLaMA, Param.=1B, Attentive probing=true2025.1274.4
InternVideo2Protocol=Zero-Shot2026.0373.1
OmniMAEType=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-H, Param.=630M, Attentive probing=true2025.1271.4
VideoMAEv2Type=Generative Pretraining, Pretrained on=Videos, Arch.=ViT-G, Param.=1.1B, Attentive probing=true2025.1271.2