This is a summary of datasets designed for image classification in deep learning. It was to help construct a 20-task combined dataset for my paper FG-AdaHAT.
| Dataset | Number of samples | Number of classes | Ave per class | Image Size |
|---|---|---|---|---|
| Arabic Handwritten Digits | 70,000 | 10 | 7,000 | 1x28x28 |
| Caltech 101 | 9,146 | 102 | 90 | Coloured, not aligned |
| Caltech 256 | 30,607 | 257 | 119 | Coloured, not aligned |
| CelebA | 202,599 | 10,177 | 20 | 3x178x218 |
| CIFAR-10 | 60,000 | 10 | 6,000 | 3x32x32 |
| CIFAR-100 | 60,000 | 100 | 600 | 3x32x32 |
| Country 211 | 63,300 | 211 | 300 | 3x256x256 |
| CUB-200-2011 | 11,788 | 200 | 59 | 3x64x64 |
| DTD | 5,640 | 47 | 120 | Coloured, not aligned, 300x300- 640x640 |
| EMNIST Balanced | 131,600 | 47 | 2800 | 1x28x28 |
| EMNIST ByClass | 814,255 | 62 | ~13000 | 1x28x28 |
| EMNIST ByMerge | 814,255 | 47 | ~17300 | 1x28x28 |
| EMNIST Digits | 280,000 | 10 | 28000 | 1x28x28 |
| EMNIST Letters | 145,600 | 26 | 5600 | 1x28x28 |
| EuroSAT | 27,000 | 10 | 2,700 | 3x64x64 |
| FaceScrub | 106,863 | 530 | 202 | Coloured, not aligned |
| FaceScrub-10 | ~10,000 | 10 | 1,000 | 3x32x32 |
| FaceScrub-20 | 20 | 3x32x32 | ||
| FaceScrub-50 | 50 | 3x32x32 | ||
| FaceScrub-100 | 100 | 3x32x32 | ||
| Fashion-MNIST | 70,000 | 10 | 7,000 | 1x28x28 |
| FER2013 | 35,887 | 7 | 5,127 | 1x48x48 |
| FGVC-Aircraft By Family | 10,200 | 70 | 146 | Coloured, not aligned |
| FGVC-Aircraft By Manufacturer | 10,200 | 41 | 249 | Coloured, not aligned |
| FGVC-Aircraft By Variant | 10,200 | 102 | 100 | Coloured, not aligned |
| Food-101 | 101,000 | 101 | 1,000 | Coloured, not aligned |
| GTSRB | 51,839 | 43 | 1,205 | Coloured, not aligned |
| ImageNet | 1,281,167 | 1,000 | 1,281 | Coloured, not aligned |
| Imagenette | 130,000 | 10 | 13,000 | Coloured, not aligned (full size) / 3x320x320 / 3x160x160 |
| iNaturalist | ~3,200,000 | 10,000 | ~320 | Coloured, not aligned |
| Kannada-MNIST | 70,000 | 10 | 7,000 | 1x28x28 |
| Kuzushiji-MNIST | 70,000 | 10 | 7,000 | 1x28x28 |
| LFW | ~13,000 | 5,749 | 2 | 3x250x250 |
| Linnaeus 5 | 8,000 | 5 | 1,600 | 3x256x256 / 3x128x128 / 3x64x64 / 3x32x32 |
| LSUN | ~10,000,000 | 10 | ~1,000,000 | Coloured, not aligned |
| MedMNIST2D BloodMNIST | 17,092 | 8 | 2,136 | 1x28x28 |
| MedMNIST2D BreastMNIST | 780 | 2 | 390 | 1x28x28 |
| MedMNIST2D ChestMNIST | 112,120 | 2 | 56,060 | 1x28x28 |
| MedMNIST2D DermaMNIST | 10,015 | 7 | 1,430 | 1x28x28 |
| MedMNIST2D OCTMNIST | 109,309 | 4 | 27,327 | 1x28x28 |
| MedMNIST2D OrganAMNIST | 58,830 | 11 | 5,348 | 1x28x28 |
| MedMNIST2D OrganCMNIST | 23,583 | 11 | 2,144 | 1x28x28 |
| MedMNIST2D OrganSMNIST | 25,211 | 11 | 2,283 | 1x28x28 |
| MedMNIST2D PathMNIST | 107,180 | 9 | 11,000 | 1x28x28 |
| MedMNIST2D PneumoniaMNIST | 5,856 | 2 | 2,928 | 1x28x28 |
| MedMNIST2D TissueMNIST | 236,386 | 8 | 29,548 | 1x28x28 |
| MNIST | 70,000 | 10 | 7,000 | 1x28x28 |
| NotMNIST | Small ~19,000, Large ~500,000 | 10 | ~1,900, ~50,000 | 1x28x28 |
| Omniglot | 32,460 | 1,623 | 20 | 1x105x105 |
| Oxford 102 Flowers | 8,189 | 102 | 80 | Coloured, not aligned |
| Oxford-IIIT Pet | 7,349 | 37 | ~200 | Coloured, not aligned |
| PCAM | 327,680 | 2 | 163,840 | 3x96x96 |
| Places365 | ~10,000,000 | 434 | ~23,000 | Coloured, not aligned |
| QMNIST | 120,000 | 10 | 12,000 | 1x28x28 |
| Rendered SST2 | 9,613 | 2 | 4,806 | 3x448x448 |
| Semeion Handwritten Digit | 1,593 | 10 | 159 | 1x16x16 |
| Sign Language MNIST | 34,627 | 24 | 1,148 | 1x28x28 |
| Stanford Cars | 16,185 | 196 | 82 | Coloured, not aligned |
| STL-10 | 13,000 + 100,000 unlabeled data | 10 | - | 3x96x96 |
| SUN397 | 108,754 | 397 | 274 | Coloured, not aligned |
| SVHN | 99,289 (without extra) | 10 | 9,929 | 3x32x32 |
| TinyImageNet | 120,000 | 200 | 600 | 3x64x64 |
| USPS | 9,298 | 10 | 930 | 1x16x16 |