Semi Supervised Image Classification On Cifar

评估指标

Percentage error

评测结果

各个模型在此基准测试上的表现结果

模型名称	Percentage error	Paper Title	Repository
GAN	15.59	Improved Techniques for Training GANs
SimMatch	3.96	SimMatch: Semi-supervised Learning with Similarity Matching
Self Meta Pseudo Labels	4.09	Self Meta Pseudo Labels: Meta Pseudo Labels Without The Teacher	-
FixMatch+DM	4.13±0.11	-	-
LiDAM	7.48	LiDAM: Semi-Supervised Learning with Localized Domain Adaptation and Iterative Matching	-
Meta Pseudo Labels (WRN-28-2)	3.89± 0.07	Meta Pseudo Labels
LaplaceNet (CNN-13)	4.99±0.08	LaplaceNet: A Hybrid Graph-Energy Neural Network for Deep Semi-Supervised Classification
ReMixMatch	5.14	ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring	-
EnAET	4.18	EnAET: A Self-Trained framework for Semi-Supervised and Supervised Learning with Ensemble Transformations
DoubleMatch	4.65±0.17	DoubleMatch: Improving Semi-Supervised Learning with Self-Supervision
UDA	5.27	Unsupervised Data Augmentation for Consistency Training
VAT	11.36	Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning
Triple-GAN-V2 (CNN-13)	10.01	Triple Generative Adversarial Networks
GLOT-DR	10.6	Global-Local Regularization Via Distributional Robustness
FlexMatch	4.19±0.01	FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling
UPS (Shake-Shake)	4.86	In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learning
SWSA	5	There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average
Diff-SySC	3.26±0.06	Diff-SySC: An Approach Using Diffusion Models for Semi-Supervised Image Classification	-
Dual Student (600)	8.89	Dual Student: Breaking the Limits of the Teacher in Semi-supervised Learning
Dash (RA, ours)	4.08±0.06	Dash: Semi-Supervised Learning with Dynamic Thresholding	-

0 of 47 row(s) selected.