As a CREODIAS user, I want to be able to access the MMEarth dataset (through the catalogue and S3/object storage) so that I can quickly execute cross-modal pre-training using MP-MAE architectures across global locations without spending weeks on data preparation. This enables me to focus on building useful applications such as label-efficient semantic segmentation and image classification, instead of solving infrastructure and data-engineering problems.
As a CREODIAS user, I want to be able to access the MMEarth dataset (through the catalogue and S3/object storage) so that I can quickly execute cross-modal pre-training using MP-MAE architectures across global locations without spending weeks on data preparation. This enables me to focus on building useful applications such as label-efficient semantic segmentation and image classification, instead of solving infrastructure and data-engineering problems.