[metascroy]

Renames the Core AI delegate manifest's packaging values so they name the embedded format instead of where it's stored.

Stacked on #24375.

Why

"inline" and "aot_compiled_inline" describe storage (embedded in the .pte), which is the same for both formats. They don't say what is actually different: a portable .aimodel versus AOT-compiled .aimodelc bundles. The runtime also used a third vocabulary, "source" / "aot", in logs and the cache identity. "source" already meant something else there: the bundle copy the runtime writes to the cache for the SDK, which happens for AOT bundles too.

Changes

Where	Before	After
Manifest packaging	"inline" / "aot_compiled_inline"	"aimodel" / "aimodelc"
Python AssetPackaging (compiler/preprocess.py)	INLINE / AOT_COMPILED_INLINE	AIMODEL / AIMODELC
Runtime manifest check (runtime/coreai_manifest.mm)	accepts the old values; AOT detected by an aot_compiled_ prefix	accepts only the new values; AOT means == "aimodelc"
Specialization log (coreai_load_coordinator.mm)	kind=source / kind=aot	kind=aimodel / kind=aimodelc
Cache identity (coreai_cache_key.mm)	"source" / "aot"	"aimodel" / "aimodelc"

Test fixtures, test labels and the README follow the new names. In the README, "source" now only means the copy written for the SDK.

Compatibility

• Old PTEs: this is a manifest format change. PTEs exported with the old names fail to load with Unsupported Core AI packaging; expected aimodel or aimodelc and must be re-exported.
• Existing caches: the cache identity changes, so existing bookmarks aren't found and each model specializes again on its first load. Old entries stay on disk unused.

Test plan

• test_preprocess.py: 25/25 passed, including the real coreai-build AOT test.
• Native: all 6 CTest targets pass (coreai_host_test, coreai_pte_export_test, coreai_swift_bridge_test, coreai_sdk_source_lifetime_test, coreai_runtime_smoke, coreai_multithread_test). The native fixtures are regenerated and carry packaging: "aimodel".
• Manual end-to-end with a standalone runner on an M4 Max:
  • Portable PTE: kind=aimodel, outputs match the reference.
  • AOT PTE: first load kind=aimodelc with exact outputs; second load restores from the cache bookmark.
  • A PTE exported before this change is rejected with the error above.
• ufmt, flake8 and clang-format are clean.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/24474

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 1 Unclassified Failure

As of commit cfdb000 with merge base 822cb60 ():

NEW FAILURE - The following job has failed:

• pull / nxp-mcuxpresso-test / linux-job (gh)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 1

UNCLASSIFIED FAILURE - DrCI could not classify the following job because the workflow did not run on the merge base. The failure may be pre-existing on trunk or introduced by this PR:

• Core AI / test-coreai / test-coreai (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 8

This comment was automatically generated by Dr. CI and updates every 15 minutes.