Skip to content

Latest commit

 

History

History
35 lines (25 loc) · 1.04 KB

satori-public-datasets.rst

File metadata and controls

35 lines (25 loc) · 1.04 KB
Raw

Satori Public Datasets

Satori use a dedicated partition of most used DataSets for machine learning and deep learning.

/nobackup/projects/public

Some of the datasets that can currently be found here:

charades/
conceptualcaptions/
facebook-co3d/
facebookresearch/
howto100m/
ImageNet/
imagenet21k_resized/
imagenet22k/
imagenet-c/
imagenet-p/
kinetics400/
kinetics600/
ProjectCodeNet/
RealEstate10k/

ImageNet dataset is stored in two forms: - original files, not processed as .tar files in the /data/ImageNet/archive folder - processed dataset, in the /data/ImageNet/ILSVRC2012. In this folder you will find the images for training (train folder) and validation (val folder)

If you want other public datasets with large general usage and interest please get in contact with the support team. They will need a URL to the downloadable dataset and will have to confirm that it actually is public and does not require any sort of per-user acknowledgements.