AI & Big Data

Hardware & Software Resources

Bridges holds a a rich set of software, containers, frameworks and environments to engage AI and Big Data research.

Containers

Bridges supports Singularity containers. You can import your own, but many are already installed on Bridges for use on the DGX-2, the Volta GPUs and the P100 GPUs.

The preinstalled containers are part of the NVIDIA GPU Container Registry. They have been built with the lastest versions of AI software and optimized by NVIDIA for the Volta and P100 nodes.

Installed containers include:

  • Caffe
  • Caffe2
  • DIGITS (NVIDIA Deep Learning GPU Training System)
  • Microsoft Cognitive Toolkit (formerly CNTK)
  • MXNet
  • PyTorch
  • TensorFlow
  • TensorFlowRT
  • TensorFlowRT Inference Server
  • Theano
  • Torch

Environments

Create your own environment with Anaconda or python virtualenv, or use one custom-built by PSC for AI applications using Anaconda and Bridges’ GPU nodes.

Custom-built environments tailored to AI, Big Data and machine learning include popular software such as TensorFlow, Theano, Keras, pandas, openCV, scikit-learn and more.

AI & Big Data software

Modern software for AI and Big Data applications are available on Bridges, including: