benchmark dataset
C13913
concept
A benchmark dataset is a standardized collection of data designed to objectively evaluate, compare, and validate the performance of algorithms, models, or systems on specific tasks.
All labels observed (7)
| Label | Occurrences |
|---|---|
| benchmark dataset canonical | 14 |
| image dataset | 11 |
| computer vision dataset | 5 |
| few-shot learning benchmark | 2 |
| image classification dataset | 2 |
| reference data set | 2 |
| meta-learning benchmark | 1 |
Instances (18)
| Instance | Via concept surface |
|---|---|
| DUCET | reference data set |
| Default Unicode Collation Element Table | reference data set |
| Modified National Institute of Standards and Technology database | image dataset |
| Fashion-MNIST | image dataset |
| KMNIST | image dataset |
| EMNIST | image dataset |
| ImageNet | image dataset |
| CIFAR-10 | image classification dataset |
| CIFAR-100 | image classification dataset |
| COCO | image dataset |
| CelebA | image dataset |
| SVHN | image dataset |
| MNIST | — |
| MSCOCO | image dataset |
| Flickr8k | — |
| Flickr30k | image dataset |
| Omniglot | — |
| miniImageNet | image dataset |