Semi Supervised Image Classification On Cifar 2

Percentage error

Résultats

Résultats de performance de divers modèles sur ce benchmark

Tableau comparatif

Nom du modèle	Percentage error
self-meta-pseudo-labels-meta-pseudo-labels	21.68
dash-semi-supervised-learning-with-dynamic	21.97±0.14
class-aware-contrastive-semi-supervised	19.32
flexmatch-boosting-semi-supervised-learning	21.90±0.15
freematch-self-adaptive-thresholding-for-semi	21.68
regularization-with-stochastic	39.19
simple-similar-pseudo-label-exploitation-for	21.89
shot-vae-semi-supervised-deep-generative	25.3
repetitive-reprediction-deep-decipher-for	32.87
laplacenet-a-hybrid-energy-neural-model-for	22.11± 0.23
simmatch-semi-supervised-learning-with	20.58
dual-student-breaking-the-limits-of-the	32.77
dp-ssl-towards-robust-semi-supervised	22.24±0.31
all-labels-are-not-created-equal-enhancing	24.45±0.12
fixmatch-simplifying-semi-supervised-learning	23.18±0.11
temporal-ensembling-for-semi-supervised	38.65
np-match-when-neural-processes-meet-semi	21.22
milking-cowmask-for-semi-supervised-image	23.07±0.30
enaet-self-trained-ensemble-autoencoding	26.93±0.21
contrastive-regularization-for-semi	21.03
fixmatch-simplifying-semi-supervised-learning	22.6
enaet-self-trained-ensemble-autoencoding	22.92
doublematch-improving-semi-supervised	21.22± 0.17
in-defense-of-pseudo-labeling-an-uncertainty-1	32
Modèle 25	20.42±0.17
semi-supervised-learning-with-self-supervised	38.7
lidam-semi-supervised-learning-with-localized	23.22