Semi Supervised Image Classification On 1

Métriques

Top 1 Accuracy

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	Top 1 Accuracy	Paper Title
PAWS (ResNet-50)	66.5%	Semi-Supervised Learning of Visual Features by Non-Parametrically Predicting View Assignments with Support Samples
Rotation	-	S4L: Self-Supervised Semi-Supervised Learning
SimCLRv2 (ResNet-152 x3, SK)	74.9%	Big Self-Supervised Models are Strong Semi-Supervised Learners
VICREG (Resnet-50)	54.8%	VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning
Exemplar (joint training)	-	S4L: Self-Supervised Semi-Supervised Learning
SEER (RegNet10B)	62.4%	Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision
SimCLRv2 distilled (ResNet-50 x2, SK)	75.9%	Big Self-Supervised Models are Strong Semi-Supervised Learners
Semi-ViT (ViT-Base)	71%	Semi-supervised Vision Transformers at Scale
Barlow Twins (ResNet-50)	55%	Barlow Twins: Self-Supervised Learning via Redundancy Reduction
DebiasPL (ResNet-50)	71.3%	Debiased Learning from Naturally Imbalanced Pseudo-Labels
SimMatchV2 (ResNet-50)	71.9%	SimMatchV2: Semi-Supervised Learning with Graph Consistency
PAWS (ResNet-50 2x)	69.6%	Semi-Supervised Learning of Visual Features by Non-Parametrically Predicting View Assignments with Support Samples
SimCLR (ResNet-50 4×)	63.0%	A Simple Framework for Contrastive Learning of Visual Representations
BYOL (ResNet-50)	53.2%	Bootstrap your own latent: A new approach to self-supervised Learning
CPC	-	Representation Learning with Contrastive Predictive Coding
SimCLRv2 distilled (ResNet-50)	73.9%	Big Self-Supervised Models are Strong Semi-Supervised Learners
SimCLR (ResNet-50 2×)	58.5%	A Simple Framework for Contrastive Learning of Visual Representations
CoMatch + EPASS (ResNet-50)	67.4%	Debiasing, calibrating, and improving Semi-supervised Learning performance via simple Ensemble Projector
Exemplar	-	S4L: Self-Supervised Semi-Supervised Learning
CoMatch (w. MoCo v2)	67.1%	CoMatch: Semi-supervised Learning with Contrastive Graph Regularization

0 of 58 row(s) selected.