Model Loading

Purpose

Loads the bundled exported model resources into a CoreAILanguageModel and opens a LanguageModelSession, so chatting runs fully on device.

As a user, I want to load the local model with one click so that I can start chatting without any network or account.

On launch the app briefly explains that the bundled model runs locally and shows a prominent Load Model button; nothing is loaded until it is pressed.
While loading, a progress indicator is shown and the UI stays responsive.
Loading reads the complete exported resource directory bundled with the app, not just the .aimodel.
After a successful load the session is prewarmed and the chat view appears.
If the resources are missing or incompatible with the CoreAILanguageModels runtime, the error view is shown instead.