STL-10 (Self-Taught Learning 10)

Introduced by Adam Coates et al. in An Analysis of Single-Layer Networks in Unsupervised Feature Learning

The STL-10 is an image dataset derived from ImageNet and popularly used to evaluate algorithms of unsupervised feature learning or self-taught learning. Besides 100,000 unlabeled images, it contains 13,000 labeled images from 10 object classes (such as birds, cats, trucks), among which 5,000 images are partitioned for training while the remaining 8,000 images for testing. All the images are color images with 96×96 pixels in size.

Source: Unsupervised Feature Learning with C-SVDDNet

Homepage

Benchmarks

Add a new result Link an existing benchmark

Task	Dataset Variant	Best Model
Image Classification	STL-10	VIT-L/16
Image Generation	STL-10	Diffusion ProjectedGAN
Image Clustering	STL-10	TEMI DINO ViT-B
Unsupervised Image Classification	STL-10	HUME
Unsupervised Anomaly Detection with Specified Settings -- 10% anomaly	STL-10	Shell-Renormalized
Out-of-Distribution Detection	STL-10	Mixup
Unsupervised Anomaly Detection with Specified Settings -- 30% anomaly	STL-10	Shell-Renormalized
Unsupervised Anomaly Detection with Specified Settings -- 20% anomaly	STL-10	Shell-Renormalized
Unsupervised Anomaly Detection with Specified Settings -- 0.1% anomaly	STL-10	LVAD
Unsupervised Anomaly Detection with Specified Settings -- 1% anomaly	STL-10	LVAD
Neural Architecture Search	STL-10	NAT-M3
Semi-Supervised Image Classification	STL-10	EnAET
Self-Supervised Learning	STL-10	ResNet50
Image Compression	STL-10	Lossyless Compressor
Unsupervised Anomaly Detection	STL-10	LVAD
Fine-Grained Image Classification	STL-10	Pre trained wide-resnet-101
Anomaly Detection	STL-10	Shell-based Anomaly