====== ML/AI tools on the POWER-CTE (P9) cluster ======

We recommend using a Conda environment, using the OPEN-CE distribution, rather than requesting (which implies difficult compiling processes) new Earth modules. Follow the steps listed below:

  mkdir /gpfs/projects/bsc32/bsc32XXX
  
where XXX are the last 3 digits of your cluster username,

  module load anaconda3

  conda config --prepend channels https://ftp.osuosl.org/pub/open-ce/current/ 

  conda create -p /gpfs/projects/bsc32/bsc32XXX/opence LIST_OF_PACKAGES 

where LIST_OF_PACKAGES is a list of packages, for example: 

  conda create -p /gpfs/projects/bsc32/bsc32XXX/opence tensorflow tensorflow-addons horovod jupyterlab xarray dask holoviews hvplot xarray scikit-learn opencv seaborn netcdf4

  conda activate /gpfs/projects/bsc32/bsc32XXX/opence
  
you can use pip once you have activated your conda environment: 

  pip install geoviews dask_jobqueue