KMeans. Includes different python and cpp implementations.
A cleaner fast python wrapper of MARIGOLD can be found here: https://github.com/AU-DIS/pyMARIGOLD
Python code includes six algorithms (mentioned in [1]): Lloyed, Elkan, Hamerly, Elkan with Hamerly, Stepwise and MARIGOLD. The codes can be executed from the notebook.
Raw .h5 files for misfit data sets have been made available through Git LFS: https://docs.github.com/en/repositories/working-with-files/managing-large-files/about-git-large-file-storage
Installation guide: https://docs.github.com/en/repositories/working-with-files/managing-large-files/installing-git-large-file-storage
They can be found in Data/ folder and converted to another format using pandas. NOTE: The provided files have not been DCT transformed. You must do so first to obtain the best performance from MARIGOLD.
If the above fails then the Git LFS is out of storage. Please contact the authors to get the data provided.
DeepGlobe Land Cover can be found online. Here is a current link: https://www.kaggle.com/datasets/balraj98/deepglobe-land-cover-classification-dataset/data