Don't keep around additional cpu copy of loaded mesh files

[Wumpf]

What

This is mainly a refactor of re_renderer's model loading pipeline, but as the title (== changelog entry) points out, a nice side-effect that arose culling unnecessary memory usage which may be important for large meshes.

@EtaLoop's attempt to add a color option to Asset3D (see #7458) made it clear that the output of the mesh loaders is really hard to work with:
Prior to this PR, they would eagerly create gpu-sided meshes and then store them alongside an optional (but always filled-out) "cpu mesh" (in essence the unpacked model file).

Now instead all model loading goes to a intermediate CpuModel which can be rather easily post processed.
Gpu resources are then created as a separate step, consuming the CpuModel (it should be trivial to create a variant that doesn't consume if we need this in the future)

Checklist

• I have read and agree to Contributor Guide and the Code of Conduct
• I've included a screenshot or gif (if applicable)
• I have tested the web demo (if applicable):
  • Using examples from latest main build: rerun.io/viewer
  • Using full set of examples from nightly build: rerun.io/viewer
• The PR title and labels are set such as to maximize their usefulness for the next release's CHANGELOG
• If applicable, add a new check to the release checklist!
• If have noted any breaking changes to the log API in CHANGELOG.md and the migration guide

• PR Build Summary
• Recent benchmark results
• Wasm size tracking

To run all checks from main, comment on the PR with @rerun-bot full-check.