docker

Intel® Explainable AI Tools Dockerfiles and images

Currently there are Dockerfiles for both Model Card Generator and Explainers and they are based on ubuntu:22.04. If you plan to build these containers from scratch, please pull the latest base first:

docker pull ubuntu:22.04

• Explainers
• Model Card Generator
• Model Card Generator UI
• Interactive
• Jupyter

Explainers

Build the image

docker compose build explainers

Check existing image:

docker images | grep -i explainers
intel/ai-tools                                      intel-ai-safety-1.2.0-explainers    9842919d01fd   1 minute ago    3.23GB

Model Card Generator

Build the image

docker compose build model_card_gen

Check existing image

docker images | grep -i mcg
intel/ai-tools                                      intel-ai-safety-1.2.0-mcg           5272f96fc69c   2 minutes ago   2.7GB

Model Card Generator UI

Build the image

docker compose build model_card_gen_ui

Check existing image

docker images | grep -i mcg-ui
intel/ai-tools                                      intel-ai-safety-1.2.0-mcg-ui        9ef60e82aea1   15 minutes ago   2.7GB

Running the UI

To run the Model Card Generator UI, you can use the docker run command.

docker run --rm -p 8051:8051 --name mcg-ui intel/ai-tools:intel-ai-safety-1.2.0-mcg-ui

Once the container is running, you can access the Model Card Generator UI by navigating to <HOST_NAME>:8051 in your web browser, where HOST_NAME is the name or IP address of the server that the container is running on.

Docker containers can run in either Interactive or Jupyter mode.

Interactive

This mode allows running the container in an interactive shell. This enables the ability to interact with the container's bash shell. Below is the command to start the container in interactive mode:

docker run --rm -it intel/ai-tools:intel-ai-safety-1.2.0-explainers bash

or

docker run --rm -it intel/ai-tools:intel-ai-safety-1.2.0-mcg bash

Jupyter

This mode launches a jupyterlab notebook server. The command below will start the jupyterlab server which can be accessed from a web browser. Each container includes jupyter kernel to enable conda environment in jupyter notebook. The port for this server is 8888 and is exposed by default when you run the container, but you can change to any port that's open on the host. In this example we are using 8887 for explainers container and 8889 for mcg container on the host.

docker run --rm -p 8887:8888 --name explainers intel/ai-tools:intel-ai-safety-1.2.0-explainers

or

docker run --rm -p 8889:8888 --name model-card-gen intel/ai-tools:intel-ai-safety-1.2.0-mcg

You can also run these containers in daemon mode:

docker run --rm -d -p 8887:8888 --name explainers intel/ai-tools:intel-ai-safety-1.2.0-explainers

or

docker run --rm -d -p 8889:8888 --name model-card-gen intel/ai-tools:intel-ai-safety-1.2.0-mcg

Finally, on your favorite browser, navigate to <HOST_NAME>:8887 where HOST_NAME is the name or IP address of the server that the container is running on. If asked for a token, review the container logs to locate the token for the Jupyter server.

docker logs -f model-card-gen

or

docker logs -f explainers

Final notes

To run the containers, you can also use docker compose this way for example:

docker compose run model_card_gen

or to run in the container in daemon mode:

docker compose run -d model_card_gen

These containers are built with intelai as default non-root container user. If you prefer to run these containers with a different user, you can build them with custom build-arg's. For example this command builds the containers with current system user along with user's id and group:

docker compose build --build-arg NON_ROOT_USER=$(id -un) --build-arg UID=(id -u) --build-arg GID=$(id -g)