Connecting Watson Machine Learning Accelerator to Watson Machine Learning

After deploying both the Watson Machine Learning service and the Watson Machine Learning Accelerator service, you can connect the two services to share training resources and use Watson Machine Learning resources with Watson Machine Learning Accelerator.

1. In WML, run the updateWMLClusterdetails.sh command line utility which allows WML to locate and use a WML Accelerator instance.
   1. Obtain the WML CASE package. Download ibm-wml-cpd-4.7.0.tgz from IBM Git.
   2. Extract the WML files from the tar package.

      tar -zxvf ibm-wml-cpd-4.7.0.tgz

   3. Navigate to the ibm-wml-cpd/inventory/wmlOperatorSetup/files directory.

      cd ibm-wml-cpd/inventory/wmlOperatorSetup/files

   4. Switch to the Cloud Pak for Data project.

      oc project cpd-instance

   5. Export the Watson Machine Learning namespace.

      export NAMESPACE=wml-namespace

      where wml-namespace is the name of the Watson Machine Learning namespace.
   6. Run the updateWMLClusterdetails.sh script.

      • sh updateWMLClusterdetails.sh wmla-ingress.wmla-namespace.svc port wml-ig wml-ig-edt internal-service-name.PROJECT_CPD_INST_OPERANDS.svc:port

        where:
        • port is the port number of the wmla-ingress service. To obtain the port number, run:

          oc get svc/wmla-ingress -n wmla-namespace

          For example:

          NAME           TYPE        ...   PORT(S)
          wmla-ingress   ClusterIP   ...   30746/TCP

        • internal-service-name:port is the internal service name and the port number. To obtain the internal service name and port number, run:

          oc get svc/internal-nginx-svc -n ${PROJECT_CPD_INST_OPERANDS}

          For example:

          NAME                 TYPE        ...   PORT(S)
          internal-nginx-svc   ClusterIP   ...   12443/TCP

        For example:

        sh updateWMLClusterdetails.sh wmla-ingress.wmla.svc 30746 wml-ig wml-ig-edt internal-nginx-svc.cpd-instance.svc:12443

2. Verify that the connection was successful by checking that the Deep Learning experiment option is available.