Deep Learning Reference Stack with Tensorflow and Intel® oneAPI Deep Neural Network Library (oneDNN)

Building Locally

We have created a set of Dockerfiles that allow you to build DLRS Tensorflow with different configurations. You will be able to pick the OS (Ubuntu or Centos), then pick the flavour, either core, which is a leaner image with just the framework, and full which includes the framework and extra features. Please refer to the table below to see all features and configurations.

	Framework	Features
Core	X
Full	X	X

Framework:

TensorFlow* 1.15.3 or
TensorFlow* 2.4.0

Features:

Transformers
Horovod
Seldon core 1.2.0

NOTE: We recommend you enable Docker Buildkit to have concurrent dependency resolution and automatic garbage collection. Docker Buildkit has been integrated in Docker since 18.06, if you have an older version, please ignore this note.

Enable Docker Buildkit (see note above)

export DOCKER_BUILDKIT=1

Building Ubuntu based DLRS

Tensorflow 1.15.*

Core:

docker build --build_arg tf_ver=tf1 -t dlrs-tensorflow-ubuntu:v0.9.0 -f Dockerfile.ubuntu --target tf_core .

Full:

docker build --build_arg tf_ver=tf1 -t dlrs-tensorflow-ubuntu:v0.9.0 -f Dockerfile.ubuntu --target tf_full.

Tensorflow 2.4.0

Core:

docker build --build_arg tf_ver=tf2 -t dlrs-tensorflow2-ubuntu:v0.9.0 -f Dockerfile.ubuntu --target tf_core .

Full:

docker build --build_arg tf_ver=tf2 -t dlrs-tensorflow2-ubuntu:v0.9.0 -f Dockerfile.ubuntu --target tf_full.

Building Centos based DLRS

Tensorflow 1.15.0

Core:

docker build --build_arg tf_ver=tf1 -t dlrs-tensorflow-centos:v0.9.0 -f Dockerfile.centos --target tf_core .

Full:

docker build --build_arg tf_ver=tf1 -t dlrs-tensorflow-centos:v0.9.0 -f Dockerfile.centos --target tf_full.

Tensorflow 2.4.0

Core:

docker build --build_arg tf_ver=tf2 -t dlrs-tensorflow2-centos:v0.9.0 -f Dockerfile.centos --target tf_core .

Full:

docker build --build_arg tf_ver=tf2 -t dlrs-tensorflow2-centos:v0.9.0 -f Dockerfile.centos --target tf_full.