jerryzeyu / hpc-ml

Singularity images for the University of Arizona High Performance Computing systems

hpc-ml: Machine-learning Singularity images for UA HPC

The recipes in this repository are designed for the University of Arizona High Performance Computing systems. They build Singularity images that include common machine learning libraries including scikit-learn, tensorflow, keras, torch, as well as the Nvidia CUDNN.

Usage

This GitHub repository is connected to Singularity Hub (https://singularity-hub.org/). Please see their documentation on interacting with Singularity images from Singularity Hub. Note that there are hard limits on how many times a container can be pulled from Singularity Hub each week, so please make sure that you always singularity pull to get a local copy, for example:

singularity pull shub://clulab/hpc-ml:centos7-python37

That will download a Singularity image named hpc-ml_centos7-python37.sif that you can then use with other Singularity commands.