Action Classification On Kinetics 600

Métriques

Top-1 Accuracy

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	Top-1 Accuracy	Paper Title	Repository
D3D+S3D-G	79.1	D3D: Distilled 3D Networks for Video Action Recognition
XViT (x16)	84.5	Space-time Mixing Attention for Video Transformer
MoViNet-A5	82.7	MoViNets: Mobile Video Networks for Efficient Video Recognition
VideoMAE V2-g	88.8	VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking
MoViNet-A2	77.5	MoViNets: Mobile Video Networks for Efficient Video Recognition
PERF-Net (distilled ResNet50-G)	82.0	PERF-Net: Pose Empowered RGB-Flow Net	-
mPLUG-2	89.8	mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video
Florence (curated FLD-900M pretrain)	87.8	Florence: A New Foundation Model for Computer Vision
MoViNet-A6	83.5	MoViNets: Mobile Video Networks for Efficient Video Recognition
S3D-G (RGB)	76.6	Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification
Modèle 11	89.7	MERLOT Reserve: Neural Script Knowledge through Vision and Language and Sound	-
UniFormer-B (ImageNet-1K)	84.8	UniFormer: Unified Transformer for Efficient Spatial-Temporal Representation Learning
LGD-3D Flow	75	Learning Spatio-Temporal Representation with Local and Global Diffusion	-
TokenLearner 16at18 w. Fuser (L/10)	86.3	TokenLearner: What Can 8 Learned Tokens Do for Images and Videos?
SlowFast 8x8 (ResNet-50)	79.9	SlowFast Networks for Video Recognition
UMT-L (ViT-L/16)	90.5	Unmasked Teacher: Towards Training-Efficient Video Foundation Models
SlowFast 16x8 (ResNet-101 + NL)	81.8	SlowFast Networks for Video Recognition
EVA	89.8%	EVA: Exploring the Limits of Masked Visual Representation Learning at Scale
I3D (RGB)	73.6	A Short Note about Kinetics-600
TubeVit-L	91.5	Rethinking Video ViTs: Sparse Video Tubes for Joint Image and Video Learning

0 of 65 row(s) selected.