Scalable deployment of cloud infrastructures for developing and processing Deep Learning models

This repository will contain the code used in my thesis project.

Infrastructure Manager

This repository contains the RADL files required to deploy the whole architecture.

Ansible Roles

The Ansible Roles that are used during deployment:

Hadoop File System

https://github.com/JJorgeDSIC/ansible-role-hadoop-1

Distributed TensorFlow Training

https://github.com/JJorgeDSIC/ansible-role-test-async

Distributed TensorFlow code & scripts

Code to run TensorFlow training in a distributed way (based on Resnet example from TF repository)

https://github.com/JJorgeDSIC/DistributedTensorFlowCodeForIM