Giters

kserve / modelmesh

Distributed Model Serving Framework

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

ModelMesh serving not working with Kserve Client APIs

abhisheksinghblr opened this issue · comments

commented

I used the KServe Client API to serve the model in modelmesh serving.

name='iris-minio'
_namespace = utils.get_default_target_namespace()
default_model_spec = V1beta1InferenceServiceSpec(predictor=V1beta1PredictorSpec(sklearn=V1beta1SKLearnSpec(storage_uri='s3://model-serving-scikit/v4/model.pkl')))
isvc = V1beta1InferenceService(api_version=constants.KSERVE_V1BETA1,kind=constants.KSERVE_KIND, metadata=client.V1ObjectMeta(name=name, namespace=_namespace, annotations={'serving.kserve.io/deploymentMode': 'ModelMesh', 'serving.kserve.io/secretKey': 'localMinIO'}), spec=default_model_spec)
KServe = KServeClient()
KServe.create(isvc)

It created in isvc

but then it does not yet ready:

  Model Status:
    Copies:
      Failed Copies:  0
    Last Failure Info:
      Message:              No ServingRuntime supports specified model type and/or protocol
      Model Revision Name:  iris-minio__isvc-b789ab6f4a
      Reason:               NoSupportingRuntime
      Time:                 2022-12-22T09:45:59Z
    States:
      Active Model State:  FailedToLoad
      Target Model State:

Please check, looks to be a bug.

commented

Closing since you opened the same issue in the main repo kserve/modelmesh-serving#293.