How to install datasets

We suggest putting all datasets under the same folder (say $DATA) to ease management and following the instructions below to organize datasets to avoid modifying the source code. The file structure looks like

$DATA/
|-- cifar10/
|-- cifar100/
|–– imagenet/
|–– caltech-101/
|–– oxford_pets/
|–– stanford_cars/

If you have some datasets already installed somewhere else, you can create symbolic links in $DATA/dataset_name that point to the original data to avoid duplicate download.

Datasets list:

CIFAR10
CIFAR100
ImageNet
Caltech101
OxfordPets
StanfordCars
Flowers102
SUN397

The instructions to prepare each dataset are detailed below. To ensure reproducibility and fair comparison for future work, we provide fixed train/val/test splits for all datasets except ImageNet where the validation set is used as test set. The fixed splits are either from the original datasets (if available) or created by us.

CIFAR10

Create a folder named cifar10/ under $DATA.
Create images/ under cifar10/.
Download the dataset from the Website and extract the training and test sets to $DATA/cifar10/images. The directory structure should look like

cifar10/
|–– images/
|   |–– train/ # contains 10 folders like apple, bus, etc.
|   |–– test/

Download the classnames.txt to $DATA/cifar10/ from this link.

CIFAR100

Create a folder named cifar100/ under $DATA.
Create images/ under cifar100/.
Download the dataset from the Website and extract the training and test sets to $DATA/cifar100/images. The directory structure should look like

cifar100/
|–– images/
|   |–– train/ # contains 100 folders like apple, bus, etc.
|   |–– test/

Download the classnames.txt to $DATA/cifar100/ from this link.

ImageNet

Create a folder named imagenet/ under $DATA.
Create images/ under imagenet/.
Download the dataset from the official website and extract the training and validation sets to $DATA/imagenet/images. The directory structure should look like
Download the BREEDS annotation in modified/ from this link

imagenet/
|–– images/
|   |–– train/ # contains 1,000 folders like n01440764, n01443537, etc.
|   |–– val/
|–– modified/

If you had downloaded the ImageNet dataset before, you can create symbolic links to map the training and validation sets to $DATA/imagenet/images.
Download the classnames.txt to $DATA/imagenet/ from this link. The class names are copied from CLIP.