kserve / kserve/open-inference-protocol

Optional model version

Open
#14 0 comments 1 reaction 0 assignees View on GitHub
Dominant language
No language data
Stars
77
Forks
15
PR merge metrics
No merged PRs in 30d

Description

The descriptions for "_Model Ready_" and "_Model Metadata_" in both the REST and gRPC specifications claim that the model version is optional. Is this true? If so, the specification should report it's optionality in more than a comment/description (`optional` in proto3 and `required: false` in openAPI). If the version is optional in the REST protocol, should there instead be another versionless route instead of eg, `'/v2/models/${MODEL_NAME}/versions/${MODEL_VERSION}/ready'`?

https://github.com/kserve/open-inference-protocol/blob/eeb043e2c95c1dd35f8b9185f5dadf44c7f10c9f/specification/protocol/open_inference_grpc.proto#L68-L70

https://github.com/kserve/open-inference-protocol/blob/eeb043e2c95c1dd35f8b9185f5dadf44c7f10c9f/specification/protocol/open_inference_grpc.proto#L98-L100

https://github.com/kserve/open-inference-protocol/blob/eeb043e2c95c1dd35f8b9185f5dadf44c7f10c9f/specification/protocol/open_inference_rest.yaml#L61

https://github.com/kserve/open-inference-protocol/blob/eeb043e2c95c1dd35f8b9185f5dadf44c7f10c9f/specification/protocol/open_inference_rest.yaml#L104

Contributor guide

No contributing guide indexed for this repository

Research direction

Read the referenced sections of specification/protocol/open_inference_grpc.proto and specification/protocol/open_inference_rest.yaml, focusing on the Model Ready and Model Metadata definitions and their versioned routes. Determine whether model versions are optional in both protocols and whether REST needs a versionless route. Done means the specification consistently records the decision in the relevant proto and OpenAPI definitions.

Written by the indexing model from the issue text.

Assessment

Tech stack
grpc, openapi
Domain
api
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.