0.0.2

update version to 0.0.2

0.0.1

New release of MLCube and MLCube runners

legacy-0.2.3

Update: This release has been tagged as legacy. Users should use mlcube.

Updates in this release

1. More configuration options for SSH runner.
2. New Kubernetes runner.
3. Minor improvements and bug fixes.

legacy-0.2.2

Update: This release has been tagged as legacy. Users should use mlcube.

Running MLBox examples.

Two example MLBoxes are available: MNIST and Hello World.

Running examples from master branch

# Clone the project
git clone https://github.com/mlperf/mlbox.git && cd ./mlbox

# Setup python environment
virtualenv -p python3.8 ./env && source ./env/bin/activate
pip install -r ./mlcommons_box/requirements.txt
export PYTHONPATH=$(pwd)/mlcommons_box:$(pwd)/runners/mlcommons_box_singularity:$(pwd)/runners/mlcommons_box_docker:$(pwd)/runners/mlcommons_box_ssh

Hello World Example

# Change directory and configure MLBox
cd ./examples/hello_world
python -m mlcommons_box_docker configure --mlbox=. --platform=platforms/docker.yaml

# Use docker runner to run two tasks
python -m mlcommons_box_docker run --mlbox=. --platform=platforms/docker.yaml --task=run/alice/hello.yaml
python -m mlcommons_box_docker run --mlbox=. --platform=platforms/docker.yaml --task=run/alice/bye.yaml

# Look at the results
cat ./workspace/chats/chat_with_alice.txt
cd ../../

MNIST Example

cd ./examples/mnist

# Configure MNIST MLBox and run two tasks - download data and model training using docker runner
python -m mlcommons_box_docker configure --mlbox=. --platform=platforms/docker.yaml
python -m mlcommons_box_docker run --mlbox=. --platform=platforms/docker.yaml --task=run/download.yaml
python -m mlcommons_box_docker run --mlbox=. --platform=platforms/docker.yaml --task=run/train.yaml

# Remove tasks' outputs
sudo rm -r ./workspace/data/ ./workspace/download_logs/ ./workspace/model/ ./workspace/train_logs/

# Do the same with singularity runner
python -m mlcommons_box_singularity configure --mlbox=. --platform=platforms/singularity.yaml
python -m mlcommons_box_singularity run --mlbox=. --platform=platforms/singularity.yaml --task=run/download.yaml
python -m mlcommons_box_singularity run --mlbox=. --platform=platforms/singularity.yaml --task=run/train.yaml

legacy-0.2.1

Update: This release has been tagged as legacy. Users should use mlcube.

legacy-0.2

Update: This release has been tagged as legacy. Users should use mlcube.

legacy-0.1

Update: This release has been tagged as legacy. Users should use mlcube.