feat(operator): Included LLM spec to CRD

[RobertSamoilescu]

What this PR does / why we need it:

This PR adds support for the llm tag in the model CRD. Similarly to the explainer tag, the llm tag allows to reference an inference model by the name. This specification is used in the PromptRuntime for the LLM Module.

Example usage:

apiVersion: mlops.seldon.io/v1alpha1
kind: Model
metadata:
  name: chat-completions
spec:
  storageUri: "gs://seldon-models/llm-runtimes/prompting-test/models/chat-completions/"
  llm:
    modelRef: tiny-llama
  requirements:
  - prompt

Which issue(s) this PR fixes:

Fixes #

Special notes for your reviewer:

[CLAassistant]

All committers have signed the CLA.

[sakoush]

In general it looks good, I left some comments for consideration.

[sakoush]

nit: I think that we should have a test case specific to LLM spec and not have it with explainer. In fact the user should not be able to set the two together and this can should be marked as invalid perhaps.

[sakoush]

LGTM!