kserve: Unable to serve on AWS S3

/kind bug

What steps did you take and what happened:

I want to serve a pytorch model using kf serving on AWS (i am following this tutorial: https://github.com/kubeflow/kfserving/tree/master/docs/samples/v1alpha2/pytorch). Here are my steps:

s3_secret.yaml

apiVersion: v1
kind: Secret
metadata:
  name: mysecret
  annotations:
     serving.kubeflow.org/s3-endpoint: s3://kfserving-samples/models/pytorch/cifar10 # replace with your s3 endpoint
     serving.kubeflow.org/s3-usehttps: "1" # by default 1, for testing with minio you need to set to 0
type: Opaque
data:
  AWS_ACCESS_KEY_ID: bWluaW8=
  AWS_SECRET_ACCESS_KEY: bWluaW8xMjM=
---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: sa
secrets:
  - name: mysecret

$ kubectl apply -f s3_secret.yaml

secret/mysecret configured
serviceaccount/sa configured

You then set the serviceAccountName on the KFService

pytorch_serve.yaml

apiVersion: "serving.kubeflow.org/v1alpha2"
kind: "InferenceService"
metadata:
  name: "pytorch-s3-cifar10-005"
spec:
  default:
    predictor:
      serviceAccountName: sa
      pytorch:
        storageUri: "s3://kfserving-samples/models/pytorch/cifar10"

$ kubectl apply -f pytorch_serve.yaml
inferenceservice.serving.kubeflow.org/pytorch-s3-cifar10-005 created

I dont get any URL generated in the output

$ kubectl get inferenceservice
NAME                     URL                                          READY   DEFAULT TRAFFIC   CANARY TRAFFIC   AGE

pytorch-s3-cifar10-005                                                False                                      6m15s

What did you expect to happen: $ kubectl get inferenceservice should have generated URL like: http://pytorch-cifar10-005.default.example.com

Anything else you would like to add: Am I missing any steps ? What should I do to make this work ?

Environment:

• Kubeflow version: 1.2
• Kubernetes version: (use kubectl version): 1.18.9
• OS (e.g. from /etc/os-release): ubuntu 18.04