feat(config): enable partial model loading by default

[Pfannkuchensack]

Summary

Flips enable_partial_loading default from false to true so low-VRAM mode is on out of the box. Updates the low-VRAM docs and generated settings reference accordingly.

Related Issues / Discussions

None

QA Instructions

• Start Invoke without setting enable_partial_loading in invokeai.yaml and confirm partial loading is active (visible in startup logs / model cache behavior).
• Verify generations still work end-to-end on a GPU that previously ran with full loading.
• Set enable_partial_loading: false in invokeai.yaml and confirm the old full-loading behavior can still be restored.
• Skim the rendered Low-VRAM mode docs — wording should reflect that partial loading is the default and explain how to opt out.

Merge Plan

Standard merge. Worth calling out in release notes / What's New since this changes default runtime behavior for all users.

Checklist

• The PR has a short but descriptive title, suitable for a changelog
• Tests added / updated (if applicable)
• ❗Changes to a redux slice have a corresponding migration
• Documentation added / updated (if applicable)
• Updated What's New copy (if doing a release after this PR)