/cliMLe

Machine Learning development for climate applications

Primary LanguagePython

cliMLe

Machine Learning workflow development for climate applications

Requirements

cliMLe uses the Keras framework.

Keras Installation instructions

For Anaconda:
  • Start with a brand new shell (no conda env activated):
  $ conda create -n tensorflow -c conda-forge -c cdat cdat
  $ source activate tensorflow
  • Linux Install
  $ pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.0.1-cp27-none-linux_x86_64.whl
  • OSX Install
  $ pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.8.0-py3-none-any.whl
  • Test tensorflow :
  $ python
  >>> import tensorflow
  >>>
  • Install Keras:
  $ pip install pillow h5py scikit-learn keras
  • Install additional packages:
  $ conda install xarray