Servable is not found by deploying tensorflow/serving within deployment manifest

I am following Autoscaling TensorFlow Model Deployments with TF Serving and Kubernetes along the related kubernetes yaml manifest on their github repo. Except that instead of GCP, I am aiming at deploying it on my local machine with the minikube node. After downloading ResNet101, the current dir looks like this:

enter image description here

By modifying the MODEL_PATH in ConfigMap from gs to the model absolute path and applying:

apiVersion: v1
kind: ConfigMap
  name: tfserving-configs
  MODEL_NAME: image_classifier
  MODEL_PATH: /home/myname/.../models/resnet_101

and applying the same suggested deployment manifest in the instruction, the pod stays in the running status and never gets ready. By logging into deployment the error is:

…: E
FileSystemStoragePathSource encountered a filesystem access error:
Could not find base path /home/myname/…/models/resnet_101 for servable
image_classifier with error NOT_FOUND: /home/myname/…/models/resnet_101 not found

I also tried to replace MODEL_PATH in configmap to the local path and reapplied it:

  MODEL_PATH: /models/resnet_101

yet again similar error.
I don’t see any reason why it is not working, unless tensorflow/serving internally can read from gs via its model_base_path argument, then I should somehow mount the model onto the pod?!