List your company

Action Recognition on HMDB51 (Top-1 and Top-5 Accuracy)

44.75Top-1 AccVLM-RLAIF
11.7320.302528.87537.4475Aug 5, 2021Apr 25, 2022Jan 14, 2023Oct 5, 2023Jun 25, 2024Mar 16, 2025Dec 5, 2025

Evaluation results

MethodLinks
VLM-RLAIFLLM Size=7B, Zero-shot=true2024.0244.7568.37
STSepPretrained=ImgNet, Params=11.5M, FLOPs×views=9.60 × 3 G, Frames=16×1282025.1241.474.7
STSep(S1-2)Pretrained=ImgNet, Params=11.3M, FLOPs×views=9.58 × 3 G, Frames=16×1282025.1240.574.3
TDBNPretrained=ImgNet, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1240.173.9
TKSPretrained=ImgNet, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.124074
RSNNPretrained=ImgNet, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1239.973.8
TcjaPretrained=ImgNet, Params=11.3M, FLOPs×views=9.49 × 3 G, Frames=16×1282025.1239.973.5
STSepPretrained=ImgNet, Params=11.5M, FLOPs×views=9.60 × 1 G, Frames=16×1282025.1239.872.4
TETPretrained=ImgNet, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1239.674
PLIFPretrained=ImgNet, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1239.173.2
LLaMA-VIDLLM Size=7B, Zero-shot=true, Reproduced=true2024.0238.8565.27
VLM-SFTLLM Size=7B, Zero-shot=true2024.0238.5862.37
Video-ChatGPTLLM Size=7B, Zero-shot=true2024.0237.163.97
Video-LLaVALLM Size=7B, Zero-shot=true2024.0236.6464.03
ER-enhanced ZSAR modelVideo=(S+Obj)+, Class=ED2021.0835.3—
E2EVideo=r(2+1)d*, Class=WN2021.0832.7—
URVideo=FV, Class=WN2021.0824.4—
TS-GCNVideo=Obj, Class=WN2021.0823.2—
ZSECOCVideo=FV, Class=WN2021.0822.6—
ASRVideo=C3D*, Class=WT2021.0821.8—
STSepPretrained=false, Params=11.5M, FLOPs×views=9.60 × 3 G, Frames=16×1282025.1221.250.9
STSepPretrained=false, Params=11.5M, FLOPs×views=9.60 × 3 G, Frames=8×1282025.1220.850.9
MTEVideo=FV, Class=WN2021.0819.7—
TDBNPretrained=false, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1219.149.5
TETPretrained=false, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1218.747.2
ESZSLVideo=FV, Class=WN2021.0818.5—
TSNPretrained=false, Params=11.3M, FLOPs×views=9.48 × 3 G, Frames=16×1282025.1217.248.2
O2AVideo=Obj, Class=WN2021.0815.6—
SJEVideo=FV, Class=WN2021.0813.3—
SVEVideo=BoW, Class=WN2021.0813—