Description of problem:

We need to update a configuration stanza for how to use Podman AI as per https://docs.redhat.com/en/documentation/migration_toolkit_for_applications/8.0/html-single/configuring_and_using_red_hat_developer_lightspeed_for_mta/index#configuring-llm-podman_configuring-llm

The configuration snippet where "base_url" is specified is broken:

Current info is broken:

  podman_mistral: &active
    provider: "ChatOpenAI"
     environment:
      OPENAI_API_KEY: "unused value"
    args:
      model: "mistral-7b-instruct-v0-2"
      base_url: "http://localhost:35841/v1"

• Need to use "baseURL"
• Need to have "baseURL" under a "configuration:" parameter we lacked in example

Here is a working snippet tested with MTA 8.0.2 and podman AI
 

 podman_granite: &active
    provider: "ChatOpenAI"
    environment:
      OPENAI_API_KEY: "unused value"
    args:
      model: "ibm-granite/granite-3.3-8b-instruct-GGUF"
      configuration:
        baseURL: "http://localhost:56885/v1"

Here is a gist of a provider-settings.yaml tested with podman AI and MTA 8.0.2:

https://gist.github.com/jwmatthews/67ff87421461fc0896ac8f8d50e05317

Note, for above I am explicitly testing with granite-3.3-8b-instruct-GGUF as I no longer see the older mistral model in the podman AI catalog.