Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

SotA/Computer Vision/Semi-Supervised Image Classification

Semi-Supervised Image Classification

80 benchmarks167 papers

Semi-supervised image classification leverages unlabelled data as well as labelled data to increase classification performance.

You may want to read some blog posts to get an overview before reading the papers and checking the leaderboards:

An overview of proxy-label approaches for semi-supervised learning - Sebastian Ruder
Semi-Supervised Learning in Computer Vision - Amit Chaudhary

<span style="color:grey; opacity: 0.6">( Image credit: Self-Supervised Semi-Supervised Learning )</span>

Benchmarks

Semi-Supervised Image Classification on CIFAR-10, 4000 Labels

Percentage error

Semi-Supervised Image Classification on cifar-100, 10000 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-10, 250 Labels

Percentage error

Semi-Supervised Image Classification on SVHN, 1000 labels

Semi-Supervised Image Classification on CIFAR-10, 40 Labels

Percentage error

Semi-Supervised Image Classification on SVHN, 250 Labels

Semi-Supervised Image Classification on CIFAR-10, 1000 Labels

Semi-Supervised Image Classification on CIFAR-100, 400 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-100, 2500 Labels

Percentage error

Semi-Supervised Image Classification on ImageNet - 1% labeled data

Top 5 Accuracy Top 1 Accuracy Number of params

Semi-Supervised Image Classification on ImageNet - 10% labeled data

Top 5 Accuracy Top 1 Accuracy Number of params

Semi-Supervised Image Classification on STL-10, 1000 Labels

Semi-Supervised Image Classification on SVHN, 500 Labels

Semi-Supervised Image Classification on CIFAR-10

All accuracy (50% Labeled)Novel accuracy (50% Labeled)Seen accuracy (50% Labeled)All accuracy (10% Labeled)Novel accuracy (10% Labeled)Seen accuracy (10% Labeled)

Semi-Supervised Image Classification on CIFAR-10, 100 Labels (OpenSet, 6/4)

Semi-Supervised Image Classification on CIFAR-10, 2000 Labels

Semi-Supervised Image Classification on CIFAR-10, 400 Labels (OpenSet, 6/4)

Semi-Supervised Image Classification on CIFAR-10, 50 Labels (OpenSet, 6/4)

Semi-Supervised Image Classification on ImageNet-100 (TEMI Split)

Novel accuracy (50% Labeled)Seen accuracy (50% Labeled)All accuracy (50% Labeled)All accuracy (10% Labeled)Novel accuracy (10% Labeled)Seen accuracy (10% Labeled)

Semi-Supervised Image Classification on STL-10, 40 Labels

Semi-Supervised Image Classification on cifar10, 250 Labels

Percentage correct

Semi-Supervised Image Classification on CIFAR-10, 100 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-100

All accuracy (10% Labeled)Novel accuracy (10% Labeled)Seen accuracy (10% Labeled)All accuracy (50% Labeled)Novel accuracy (50% Labeled)Seen accuracy (50% Labeled)

Semi-Supervised Image Classification on STL-10

Semi-Supervised Image Classification on CIFAR-10 (250 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on CIFAR-10 (4000 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on CIFAR-10, 30 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-10, 80 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-100 (10000 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on CIFAR-100 (250 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on CIFAR-100 (400 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on CIFAR-100, 5000Labels

Percentage correct

Semi-Supervised Image Classification on DeepWeeds, 99 Labels

Percentage error

Semi-Supervised Image Classification on EuroSAT, 100 Labels

Percentage error

Semi-Supervised Image Classification on EuroSAT, 20 Labels

Percentage error

Semi-Supervised Image Classification on Imagenette, 100 Labels

Percentage error

Semi-Supervised Image Classification on Imagenette, 20 Labels

Percentage error

Semi-Supervised Image Classification on Mini-ImageNet, 4000 Labels

Semi-Supervised Image Classification on STL-10 (1000 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on SVHN (1000 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on SVHN (250 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on SVHN (40 Labels, ImageNet-100 Unlabeled)

Semi-Supervised Image Classification on SVHN, 40 Labels

Percentage error

Semi-Supervised Image Classification on cifar-10, 10 Labels

Accuracy (Test)

Semi-Supervised Image Classification on CIFAR-10, 20 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-10, 500 Labels

Semi-Supervised Image Classification on CIFAR-100, 1000 Labels

Percentage correct

Semi-Supervised Image Classification on CIFAR-100, 200 Labels

Percentage error

Semi-Supervised Image Classification on CIFAR-100, 4000 Labels

Semi-Supervised Image Classification on CIFAR-100, 5000 Labels

Semi-Supervised Image Classification on Mini-ImageNet, 1000 Labels

Semi-Supervised Image Classification on Mini-ImageNet, 10000 Labels

Semi-Supervised Image Classification on STL-10, 5000 Labels

Semi-Supervised Image Classification on SVHN, 2000 Labels

Semi-Supervised Image Classification on SVHN, 4000 Labels

Semi-Supervised Image Classification on Salinas

Overall Accuracy

Semi-Supervised Image Classification on Caltech-101

Semi-Supervised Image Classification on Caltech-101, 202 Labels

Semi-Supervised Image Classification on Caltech-256

Semi-Supervised Image Classification on Caltech-256, 1024 Labels

Semi-Supervised Image Classification on ImageNet - 0.2% labeled data

ImageNet Top-1 Accuracy