Skip to content
This repository was archived by the owner on Jul 27, 2026. It is now read-only.

docker.md

Latest commit

 

History

History
57 lines (46 loc) · 2.54 KB

docker.md

File metadata and controls

57 lines (46 loc) · 2.54 KB
Raw

Create IBMFL Docker Container Image

Install Docker engine by following instructions from official Docker webpage: https://docs.docker.com/engine/install

Next, use Dockerfile provided in the repository to create a container image. The -t ibmfl flag can be changed to whichever image name is desired.

docker build -t ibmfl .

Just like installing the wheel file, by default, this will not install any additional machine learning libraries. You must specify the desired model training backend library by using the --build-arg flag to specify a BACKEND argument. The following backends are supported:

# Build image with Scikit-learn backend
docker build --build-arg BACKEND=sklearn -t ibmfl .
# Build image with PyTorch backend
docker build --build-arg BACKEND=pytorch -t ibmfl .
# Build image with Keras (TensorFlow v1) backend
docker build --build-arg BACKEND=keras -t ibmfl .
# Build image with TensorFlow v2 backend
docker build --build-arg BACKEND=tf -t ibmfl .
# Build image with RLlib backend
docker build --build-arg BACKEND=rllib -t ibmfl .

You can also use multiple backends using a comma separated list. For example:

# Build image with Scikit-learn and Keras (TensorFlow v1) backend
docker build --build-arg BACKEND=sklearn,keras -t ibmfl
# Build image with PyTorch and TensorFlow v2 backend
docker build --build-arg BACKEND=pytorch,tf -t ibmfl
# Build image with Scikit-learn, TensorFlow v2, and RLlib backend
docker build --build-arg BACKEND=sklearn,tf,rllib -t ibmfl

To build the docker with crypto using fully homomorphic encryption (HE), add crypto to the BACKEND list. Typically this is paired with other backends.

# Build image with only crypto and no backend
docker build --build-arg BACKEND=crypto -t ibmfl
# Build image with Keras (TensorFlow v1) backend and crypto
docker build --build-arg BACKEND=keras,crypto -t ibmfl
# build image with PyTorch and TensorFlow v2 backend and crypto
docker build --build-arg BACKEND=pytorch,tf,crypto -t ibmfl

To build the docker to be run using OpenShift, add openshift to the BACKEND list. This must be paired with other backends and can optionally be paired with crypto.

# Build image with Keras (TensorFlow v1) backend for OpenShift
docker build --build-arg BACKEND=keras,openshift -t ibmfl
# Build image with Keras (TensorFlow v1) backend and crypto for OpenShift
docker build --build-arg BACKEND=keras,crypto,openshift -t ibmfl
# build image with PyTorch and TensorFlow v2 backend for OpenShift
docker build --build-arg BACKEND=pytorch,tf,openshift -t ibmfl