Semi Supervised Image Classification On Cifar 2

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

		Paper Title	Repository
Ⅱ-Model	39.19	Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning	-
SESEMI SSL (ConvNet)	38.7	Exploring Self-Supervised Regularization for Supervised and Semi-Supervised Learning
Temporal ensembling	38.65	Temporal Ensembling for Semi-Supervised Learning
R2-D2 (CNN-13)	32.87	Repetitive Reprediction Deep Decipher for Semi-Supervised Learning
Dual Student (480)	32.77	Dual Student: Breaking the Limits of the Teacher in Semi-supervised Learning
UPS (CNN-13)	32	In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learning
EnAET (WRN-28-2)	26.93±0.21	EnAET: A Self-Trained framework for Semi-Supervised and Supervised Learning with Ensemble Transformations
SHOT-VAE	25.3	SHOT-VAE: Semi-supervised Deep Generative Models With Label-aware ELBO Approximations
SemCo (μ=7)	24.45±0.12	All Labels Are Not Created Equal: Enhancing Semi-supervision via Label Grouping and Co-training
LiDAM	23.22	LiDAM: Semi-Supervised Learning with Localized Domain Adaptation and Iterative Matching	-
FixMatch (CTA, WRN-28-8)	23.18±0.11	FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
CowMix (WRN-28-96x2d)	23.07±0.30	Milking CowMask for Semi-Supervised Image Classification
EnAET (WRN-28-2-Large)	22.92	EnAET: A Self-Trained framework for Semi-Supervised and Supervised Learning with Ensemble Transformations
FixMatch (RA, WRN-28-8)	22.6	FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
DP-SSL	22.24±0.31	DP-SSL: Towards Robust Semi-supervised Learning with A Few Labeled Samples	-
LaplaceNet (WRN-28-8)	22.11± 0.23	LaplaceNet: A Hybrid Graph-Energy Neural Network for Deep Semi-Supervised Classification
Dash (RA, WRN-28-8)	21.97±0.14	Dash: Semi-Supervised Learning with Dynamic Thresholding	-
FlexMatch	21.90±0.15	FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling
SimPLE (WRN-28-8)	21.89	SimPLE: Similar Pseudo Label Exploitation for Semi-Supervised Classification
SMPL (WRN-28-8)	21.68	Self Meta Pseudo Labels: Meta Pseudo Labels Without The Teacher	-

0 of 27 row(s) selected.