Provide default model in local NIM mode

[raspawar]

New implementation:

This can be simplified by using the first available model if none is provided, e.g.

llm = ChatNVIDIA(base_url="http://localhost:1234/v1") llm.chat(...)

llm._client.model => default model => first available model in local NIM

Fixes:

• Display of the lora variants in local NIM mode
• _lint.yml workflow update for the failing lint checks