kserve validate_model

Validate the proper deployment of a KServe model

Warning:

This command requires grpcurl to be available in the PATH.

Parameters

inference_service_names

• A list of names of the inference service to validate

query_count

• Number of query to perform

runtime

• Name of the runtime used (standalone-tgis or vllm)

model_id

• The model-id to pass to the inference service

• default value: not-used

namespace

• The namespace in which the Serving stack was deployed. If empty, use the current project.

raw_deployment

• If True, do not try to configure anything related to Serverless. Works only in-cluster at the moment.

method

• The gRPC method to call #TODO remove?

proto

• If not empty, the proto file to pass to grpcurl