mnist

MNIST MLCube

Create and initialize python environment

virtualenv -p python3 ./env && source ./env/bin/activate && pip install mlcube-docker mlcube-singularity mlcube-ssh

Clone MLCube examples and go to MNIST root directory

git clone https://github.com/mlperf/mlcube_examples.git && cd ./mlcube_examples/mnist

Run MNIST MLCube on a local machine with Docker runner

# Configure MNIST MLCube
mlcube_docker configure --mlcube=. --platform=platforms/docker.yaml

# Run MNIST training tasks: download data and train the model
mlcube_docker run --mlcube=. --platform=platforms/docker.yaml --task=run/download.yaml
mlcube_docker run --mlcube=. --platform=platforms/docker.yaml --task=run/train.yaml

Go to workspace/ directory and study its content. Then:

sudo rm -r ./workspace/data ./workspace/download_logs ./workspace/model ./workspace/train_logs   

Run MNIST MLCube on a local machine with Singularity runner

# Configure MNIST MLCube
mlcube_singularity configure --mlcube=. --platform=platforms/singularity.yaml

# Run MNIST training tasks: download data and train the model
mlcube_singularity run --mlcube=. --platform=platforms/singularity.yaml --task=run/download.yaml
mlcube_singularity run --mlcube=. --platform=platforms/singularity.yaml --task=run/train.yaml

Go to workspace/ directory and study its content. Then:

sudo rm -r ./workspace/data ./workspace/download_logs ./workspace/model ./workspace/train_logs   

Run MNIST MLCube on a remote machine with SSH runner

Setup passwordless access to a remote machine. Create and/or update your SSH configuration file (~/.ssh/config). Create an alias for your remote machine. This will enable access for tools like ssh, rsync and scp using mlcube-remote name instead of actual name or IP address.

Host mlcube-remote
    HostName {{IP_ADDRESS}}
    User {{USER_NAME}}
    IdentityFile {{PATH_TO_IDENTITY_FILE}}

Remove results of previous runs. Remove all directories in workspace/ except workspace/parameters.

# Configure MNIST MLCube
mlcube_ssh configure --mlcube=. --platform=platforms/ssh.yaml

# Run MNIST training tasks: download data and train the model
mlcube_ssh run --mlcube=. --platform=platforms/ssh.yaml --task=run/download.yaml
mlcube_ssh run --mlcube=. --platform=platforms/ssh.yaml --task=run/train.yaml

Go to workspace/ directory and study its content. Then:

sudo rm -r ./workspace/data ./workspace/download_logs ./workspace/model ./workspace/train_logs