kubeflow install

# Create a namespace for kubeflow deployment
NAMESPACE=kubeflow
kubectl create namespace ${NAMESPACE}

# Initialize a ksonnet app. Set the namespace for it's default environment.
APP_NAME=kf-j2m
ks init ${APP_NAME}
cd ${APP_NAME}
ks env set default --namespace ${NAMESPACE}

# Install Kubeflow components
ks registry add kubeflow github.com/kubeflow/kubeflow/tree/master/kubeflow
ks pkg install kubeflow/core
ks pkg install kubeflow/tf-serving
ks pkg install kubeflow/tf-job

ks env add nocloud
KF_ENV=nocloud
ks env set ${KF_ENV} --namespace ${NAMESPACE}

# Create templates for core components
# ks generate kubeflow-core kubeflow-core
ks generate core kubeflow-core --name=kubeflow-core

ks env add nocloud
KF_ENV=nocloud
ks env set ${KF_ENV} --namespace ${NAMESPACE}

# If your cluster is running on Azure you will need to set the cloud parameter.
# If the cluster was created with AKS or ACS choose aks, it if was created
# with acs-engine, choose acsengine
# PLATFORM=<aks|acsengine>
# ks param set kubeflow-core cloud ${PLATFORM}

# Enable collection of anonymous usage metrics
# Skip this step if you don't want to enable collection.
# ks param set kubeflow-core reportUsage true
# ks param set kubeflow-core usageId $(uuidgen)

# Deploy Kubeflow
ks apply ${KF_ENV} -c kubeflow-core