List your company
ResearchFormerly WizWand

Video Classification on UCF-101 (test)

98.1Top-1 AccUniFormer
65.8674.2382.690.97Aug 1, 2019Mar 10, 2020Oct 18, 2020May 29, 2021Jan 6, 2022Aug 16, 2022Mar 27, 2023
Updated 8mo ago

Evaluation results

MethodLinks
UniFormerPretrain=IN1K+K400, Backbone=UniFormer-S2022.0198.1
I3Duse 3D-Conv=true2019.0898
TDNPretrain=IN1K+K400, Backbone=ResNet502022.0197.4
TEAPretrain=IN1K+K400, Backbone=ResNet502022.0196.9
R(2+1)DPretrain=K400, Backbone=ResNet342022.0196.8
VidTrPretrain=IN21K+K400, Backbone=ViT-B2022.0196.6
CMA_iter1-Suse 3D-Conv=false2019.0896.5
STMPretrain=IN1K+K400, Backbone=ResNet502022.0196.2
CT-NetPretrain=IN1K+K400, Backbone=ResNet502022.0196.2
I3DPretrain=IN1K+K400, Backbone=InceptionV22022.0195.8
ResNet50-Suse 3D-Conv=false2019.0895.5
CMA_iter1-Ruse 3D-Conv=false2019.0895.3
ECOuse 3D-Conv=true2019.0894.8
Attention Clusteruse 3D-Conv=false2019.0894.6
TSMPretrain=IN1K+K400, Backbone=ResNet502022.0194.5
ARTNetuse 3D-Conv=true2019.0894.3
TSNuse 3D-Conv=false2019.0894
ResNet50-Fuse 3D-Conv=false2019.0892.4
TSNPretrain=IN1K+K400, Backbone=InceptionV22022.0191.1
ResNet50-Ruse 3D-Conv=false2019.0890.9
Two-streamuse 3D-Conv=false2019.0888
C3DPretrain=Sports-1M, Backbone=ResNet182022.0185.8
EVA-02-CLIPZero-shot=true, Backbone=ViT-E/14+, Training Scale=Enlarged2023.0383.1
EVA-02-CLIPZero-shot=true, Backbone=ViT-E/14, Training Scale=Enlarged2023.0382.8
Open CLIPZero-shot=true, Backbone=ViT-G/14, Training Scale=Enlarged2023.0380.5
EVA-01-CLIPZero-shot=true, Backbone=ViT-g/14+, Training Scale=Enlarged2023.0378.9
EVA-02-CLIPZero-shot=true, Backbone=ViT-L/14+, Training Scale=Enlarged2023.0378.6
Open CLIP-Hzero-shot evaluation protocol=True2022.1178.2
Open CLIPZero-shot=true, Backbone=ViT-H/14, Training Scale=Enlarged2023.0378.2
OpenAI CLIPZero-shot=true, Backbone=ViT-L/14+, Training Scale=Enlarged2023.0378.1
Open CLIPZero-shot=true, Backbone=ViT-g/14, Training Scale=Enlarged2023.0377.8
Open CLIP-gzero-shot evaluation protocol=True2022.1177.7
EVA-02-CLIPZero-shot=true, Backbone=ViT-L/14, Training Scale=Standard2023.0376.8
OpenAI CLIP-Lzero-shot evaluation protocol=True2022.1176.4
OpenAI CLIPZero-shot=true, Backbone=ViT-L/14, Training Scale=Standard2023.0376.4
EVA CLIP-gzero-shot evaluation protocol=True2022.1176.1
EVA-01-CLIPZero-shot=true, Backbone=ViT-g/14, Training Scale=Enlarged2023.0376
Open CLIPZero-shot=true, Backbone=ViT-L/14, Training Scale=Standard2023.0373.2
EVA-02-CLIPZero-shot=true, Backbone=ViT-B/16, Training Scale=Standard2023.0368.6
Open CLIPZero-shot=true, Backbone=ViT-B/16, Training Scale=Standard2023.0367.5
OpenAI CLIPZero-shot=true, Backbone=ViT-B/16, Training Scale=Standard2023.0367.1