Deep Learning using CNTK, Caffe, Keras +Theano,Torch, Tensorflow on Docker with Microsoft Azure Batch Shipyard

[Lee Stott]

First published on MSDN on Feb 13, 2017

Learn how to start submitting Deep neural Network training jobs using Azure N series GPU running Ubuntu on Dockers in Azure by using Azure Batch to schedule the jobs to your GPU compute clusters.

You can also deploy your Deep Neutral Network tools and libraries, on preconfigured Linux-based cluster via Docker.

This is a great way for developers to get going quickly in the world of Azure and artificial intelligence.

Get started: https://github.com/Azure/batch-shipyard

Azure Batch Recipes VMs https://github.com/Azure/batch-shipyard/tree/master/recipes

Here are some of premade Azure GPU deep Learning shipyard recipes which are ready to be deployed

Microsoft Cognitive Toolkit
CNTK-CPU-Infiniband-IntelMPI
CNTK-CPU-OpenMPI
CNTK-GPU-OpenMPI Infiniband

Caffe
Caffe-CPU
Caffe-GPU Infiniband

FFMpeg
FFmpeg-GPU

HPC
HPCG-Infiniband-IntelMPI
HPLinpack-Infiniband-IntelMPI

Keras+Theano
Keras+Theano-CPU
Keras+Theano-GPU Infiniband

MXNet
MXNet-CPU
MXNet-GPU Infiniband

NAMD
NAMD-GPU Infiniband
NAMD-Infiniband-IntelMPI
NAMD-TCP

OpenFOAM
OpenFOAM-Infiniband-IntelMPI Infiniband
OpenFOAM-TCP-OpenMPI

TensorFlow
TensorFlow-CPU
TensorFlow-Distributed Infiniband
TensorFlow-GPU Infiniband

Torch
Torch-CPU
Torch-GPU Infiniband