Installations

DeepHyper is a Python package containing different modules that can be installed in differents ways. The following sections describe the different ways to install DeepHyper. The DeepHyper package itself is a pure Python package and therefore does not need “compilation”. However, some of the (optional) modules of DeepHyper require external libraries to be installed which can require compilation when pre-built binaries are not available (or not optimized) for your target system (e.g., Tensorflow or Numpy).

Generic Installations

• conda
• docker
• jupyter
• pip (default)
• spack

High Performance Computing Centers

• Argonne Leadership Computing Facility (ALCF)
• National Energy Research Scientific Computing (NERSC)
• Oak Ridge Leadership Computing Facility (OLCF)