Skip to content
This repository has been archived by the owner on Jul 18, 2024. It is now read-only.

Latest commit

 

History

History

Dockerfile-ubuntu

README.md

Build Dockerfile for Intel® End-to-End AI Optimization Kit

How to build it

By default, you can build e2eAIOK docker image with latest nightly-build e2eAIOK PyPI distributions.

$ cd Dockerfile-ubuntu
$ docker build -t e2eaiok-tensorflow . -f DockerfileTensorflow
$ docker build -t e2eaiok-pytorch . -f DockerfilePytorch
$ docker build -t e2eaiok-pytorch112 . -f DockerfilePytorch112

If you need http and https proxy to build docker image:

$ cd Dockerfile-ubuntu
$ docker build -t e2eaiok-tensorflow . -f DockerfileTensorflow --build-arg http_proxy=http://proxy-ip:proxy-port --build-arg https_proxy=http://proxy-ip:proxy-port
$ docker build -t e2eaiok-pytorch . -f DockerfilePytorch --build-arg http_proxy=http://proxy-ip:proxy-port --build-arg https_proxy=http://proxy-ip:proxy-port
$ docker build -t e2eaiok-pytorch112 . -f DockerfilePytorch112 --build-arg http_proxy=http://proxy-ip:proxy-port --build-arg https_proxy=http://proxy-ip:proxy-port

How to use the image.

Option 1: To start a notebook directly with a specified port(e.g. 12888).

docker run -it --rm -p 12888:12888 e2eaiok-tensorflow:latest
docker run -it --rm -p 12888:12888 e2eaiok-pytorch:latest
docker run -it --rm -p 12888:12888 e2eaiok-pytorch112:latest

Option 2: To start a cluster via one-click script

git clone https://github.com/intel/e2eAIOK.git
cd e2eAIOK
git submodule update --init --recursive

Option 2.1: Start a cluster w/o jupyter notebook service enabled

python scripts/start_e2eaiok_docker.py --backend [tensorflow, pytorch, pytorch112] --dataset_path ../ --workers host1, host2, host3, host4 --proxy "http://addr:ip"

Option 2.2: Start a cluster w/ jupyter notebook service enabled

python scripts/start_e2eaiok_docker.py --backend [tensorflow, pytorch, pytorch112] --dataset_path ../ --workers host1, host2, host3, host4 --proxy "http://addr:ip" --jupyter_mode

Versions and Components

DockerfilePytorch

  • PyTorch 1.5
  • Intel® Extension for Pytorch 0.2, 1.12.x
  • 3.7.16

DockerfilePytorch112

  • PyTorch 1.12
  • Intel® Extension for Pytorch 1.12.x
  • Python 3.9.15

DockerfileTensorflow

  • TensorFlow 2.10.0
  • Intel® Extension for TensorFlow 2.10.x
  • Horovod 0.26
  • Python 3.9.12