Action Classification On Kinetics 700

Metrics

Top-1 Accuracy

Top-5 Accuracy

Results

Performance results of various models on this benchmark

Model Name	Top-1 Accuracy	Top-5 Accuracy	Paper Title	Repository
SRTG r(2+1)d-34	49.43	73.23	Learn to cycle: Time-consistent feature discovery for action recognition	-
MViTv2-B	76.6	93.2	MViTv2: Improved Multiscale Vision Transformers for Classification and Detection	-
SRTG r3d-50	53.52	74.17	Learn to cycle: Time-consistent feature discovery for action recognition	-
MoViNet-A1	63.5	-	MoViNets: Mobile Video Networks for Efficient Video Recognition	-
MoViNet-A2	66.7	-	MoViNets: Mobile Video Networks for Efficient Video Recognition	-
MoViNet-A3	68.0	-	MoViNets: Mobile Video Networks for Efficient Video Recognition	-
VidTr-M	69.5	88.3	VidTr: Video Transformer Without Convolutions	-
InternVideo-T	84.0	-	InternVideo: General Video Foundation Models via Generative and Discriminative Learning	-
EVA	82.9%	-	EVA: Exploring the Limits of Masked Visual Representation Learning at Scale	-
MoViNet-A4	70.7	-	MoViNets: Mobile Video Networks for Efficient Video Recognition	-
UniFormerV2-L	82.7	96.2	UniFormerV2: Spatiotemporal Learning by Arming Image ViTs with Video UniFormer
MaskFeat (no extra data, MViT-L)	80.4	95.7	Masked Feature Prediction for Self-Supervised Visual Pre-Training	-
InternVideo2-1B	85.4	-	InternVideo2: Scaling Foundation Models for Multimodal Video Understanding	-
SRTG r3d-101	56.46	76.82	Learn to cycle: Time-consistent feature discovery for action recognition	-
VidTr-L	70.2	89	VidTr: Video Transformer Without Convolutions	-
SRTG r3d-34	49.15	72.68	Learn to cycle: Time-consistent feature discovery for action recognition	-
mPLUG-2	80.4	94.9	mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video	-
UMT-L (ViT-L/16)	83.6	96.7	Unmasked Teacher: Towards Training-Efficient Video Foundation Models	-
CoVeR (JFT-3B)	79.8	94.9	Co-training Transformer with Videos and Images Improves Action Recognition	-
MViTv2-L (ImageNet-21k pretrain)	79.4	94.9	MViTv2: Improved Multiscale Vision Transformers for Classification and Detection	-

0 of 36 row(s) selected.