Skip to content

Namespace immutable constants per method in to_native (#23891) - #23891

Open
JakeStevens wants to merge 5 commits into
pytorch:mainfrom
JakeStevens:export-D125294428
Open

Namespace immutable constants per method in to_native (#23891)#23891
JakeStevens wants to merge 5 commits into
pytorch:mainfrom
JakeStevens:export-D125294428

Conversation

@JakeStevens

@JakeStevens JakeStevens commented Oct 15, 2026 •
edited by meta-codesync Bot
Loading

Copy link
Copy Markdown
Contributor

Summary:

Exported methods can reuse a constant name with different immutable tensors, causing to_native to reject the combined program. Namespace every immutable constant as <method>/<key>, including single-method programs, and update its graph references, including quantization metadata and nested graphs. Identical tensors still deduplicate when saving the .ptn.

Keys mutated by any method retain their original names throughout the program, preserving shared state for readers and writers. Reject differing values for shared mutable keys and collisions between final names.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428

@pytorch-bot

pytorch-bot Bot commented Oct 15, 2026 •
edited
Loading

Copy link
Copy Markdown

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/23891

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

❌ 1 Unclassified Failure

As of commit 27fff9b with merge base c45d76f (image):

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:

  • Test Native Engines / native-ggml (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
    ModuleNotFoundError: No module named 'flatbuffers'

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

@meta-cla meta-cla Bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Oct 15, 2026
@meta-codesync

meta-codesync Bot commented Oct 15, 2026

Copy link
Copy Markdown
Contributor

@JakeStevens has exported this pull request. If you are a Meta employee, you can view the originating Diff in D125294428.

@github-actions

github-actions Bot commented Oct 15, 2026

Copy link
Copy Markdown

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

@meta-codesync meta-codesync Bot changed the title Namespace colliding lifted tensor constants in to_native Namespace colliding lifted tensor constants in to_native (#23891) Oct 15, 2026
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 15, 2026
@JakeStevens @facebook-github-bot
Namespace colliding lifted tensor constants in to_native (pytorch#23891)
7567aa6 
Summary:

Edge lowering mints `_lifted_tensor_constant{i}` names independently
per method, so unrelated scalars (e.g. an fp32 attention scale and an
int64 zero) collide under one data key when packaging multiple
methods, and `to_native` rejected the package. Namespace only those
auto-generated keys per method (`<method>/<key>`) in both the merged
graph blob (new `rename_data_keys` in `graph_serialize.py`, walking
the same references as `collect_data_keys`) and the constant dicts;
identical tensors stay shared, mutated keys are never renamed, and
genuine parameter/buffer fqn collisions still raise.

Differential Revision: D125294428
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 15, 2026
@JakeStevens @meta-codesync
Namespace colliding lifted tensor constants in to_native (pytorch#23891)
91fd717 
Summary:
Pull Request resolved: pytorch#23891

Edge lowering mints `_lifted_tensor_constant{i}` names independently
per method, so unrelated scalars (e.g. an fp32 attention scale and an
int64 zero) collide under one data key when packaging multiple
methods, and `to_native` rejected the package. Namespace only those
auto-generated keys per method (`<method>/<key>`) in both the merged
graph blob (new `rename_data_keys` in `graph_serialize.py`, walking
the same references as `collect_data_keys`) and the constant dicts;
identical tensors stay shared, mutated keys are never renamed, and
genuine parameter/buffer fqn collisions still raise.

Differential Revision: D125294428
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 15, 2026
@JakeStevens @facebook-github-bot
Namespace colliding lifted tensor constants in to_native (pytorch#23891)
75bf79f 
Summary:

Edge lowering mints `_lifted_tensor_constant{i}` names independently
per method, so unrelated scalars (e.g. an fp32 attention scale and an
int64 zero) collide under one data key when packaging multiple
methods, and `to_native` rejected the package. Namespace only those
auto-generated keys per method (`<method>/<key>`) in both the merged
graph blob (new `rename_data_keys` in `graph_serialize.py`, walking
the same references as `collect_data_keys`) and the constant dicts;
identical tensors stay shared, mutated keys are never renamed, and
genuine parameter/buffer fqn collisions still raise.

Differential Revision: D125294428
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 15, 2026
@JakeStevens @meta-codesync
Namespace colliding lifted tensor constants in to_native (pytorch#23891)
bd4eaab 
Summary:
Pull Request resolved: pytorch#23891

Edge lowering mints `_lifted_tensor_constant{i}` names independently
per method, so unrelated scalars (e.g. an fp32 attention scale and an
int64 zero) collide under one data key when packaging multiple
methods, and `to_native` rejected the package. Namespace only those
auto-generated keys per method (`<method>/<key>`) in both the merged
graph blob (new `rename_data_keys` in `graph_serialize.py`, walking
the same references as `collect_data_keys`) and the constant dicts;
identical tensors stay shared, mutated keys are never renamed, and
genuine parameter/buffer fqn collisions still raise.

Differential Revision: D125294428
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 15, 2026
@JakeStevens @meta-codesync
Namespace lifted tensor constants per method in to_native (pytorch#…
1011d1d 
…23891)

Summary:
Pull Request resolved: pytorch#23891

Edge lowering generates `_lifted_tensor_constant{i}` keys independently for each method. When methods use different scalar values under the same generated key, `to_native` rejects the combined program.

Always namespace generated immutable lifted constants as `<method>/<key>`, including single-method programs. Update the constant dictionaries and graph references together through a shared `map_data_keys` traversal, including quantization metadata and nested graphs. Equal tensors under different keys still deduplicate when saving the `.ptn`. Mutable keys keep their existing names, parameter/buffer conflicts continue to raise, and explicit keys occupying a generated namespace are rejected.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428
@meta-codesync meta-codesync Bot changed the title Namespace colliding lifted tensor constants in to_native (#23891) Namespace lifted tensor constants per method in to_native (#23891) Oct 15, 2026
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 15, 2026
@JakeStevens @facebook-github-bot
Namespace lifted tensor constants per method in to_native (pytorch#…
0bac518 
…23891)

Summary:

Edge lowering generates `_lifted_tensor_constant{i}` keys independently for each method. When methods use different scalar values under the same generated key, `to_native` rejects the combined program.

Always namespace generated immutable lifted constants as `<method>/<key>`, including single-method programs. Update the constant dictionaries and graph references together through a shared `map_data_keys` traversal, including quantization metadata and nested graphs. Equal tensors under different keys still deduplicate when saving the `.ptn`. Mutable keys keep their existing names, parameter/buffer conflicts continue to raise, and explicit keys occupying a generated namespace are rejected.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428
digantdesai
digantdesai requested changes Oct 16, 2026
View reviewed changes

@digantdesai digantdesai left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Review automatically exported from Phabricator review in Meta.

@JakeStevens @meta-codesync
Pin llama.cpp for the native GGML engine
0641030 
Summary:
Add `ggml-org/llama.cpp` under `backends/native/third-party`, pinned to `620d2225cc436ab0b001e3200e590e3b286ef9ab`, matching the internal dependency. The full submodule retains the upstream license and copyright notices.

Authored with AI assistance from Codex.

Differential Revision: D125694754
@JakeStevens @meta-codesync
Add CMake targets for the native Vulkan engine and `run_classificatio…
90ec82f 
…n` runner (pytorch#23811)

Summary:
Pull Request resolved: pytorch#23811

Add engine-independent CMake libraries for the native graph, PTN package reader, program runtime, engine interface, and standalone API. `EXECUTORCH_BUILD_NATIVE` enables the common libraries; `EXECUTORCH_BUILD_NATIVE_VULKAN` also enables `native_vulkan_engine`.

`EXECUTORCH_BUILD_NATIVE_VISION_RUNNER` (requires `EXECUTORCH_BUILD_NATIVE_VULKAN`) additionally builds the `examples/vision` classification runner as `run_classification` plus its `native_runner_utils` library. The runner links `native_vulkan_engine` and gflags and installs to `bin/`. Its existing `test_runner_utils.cpp` coverage runs as `native_runner_utils_test` under `EXECUTORCH_BUILD_TESTS`.

Authored with Codex and Muse Code.

Differential Revision: D125084367
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 16, 2026
@JakeStevens @meta-codesync
Namespace immutable constants per method in to_native (pytorch#23891)
97051f4 
Summary:
Pull Request resolved: pytorch#23891

Exported methods can reuse a constant name with different immutable tensors, causing `to_native` to reject the combined program. Namespace every immutable constant as `<method>/<key>`, including single-method programs, and update its graph references, including quantization metadata and nested graphs. Identical tensors still deduplicate when saving the `.ptn`.

Keys mutated by any method retain their original names throughout the program, preserving shared state for readers and writers. Reject differing values for shared mutable keys and collisions between final names.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428
@meta-codesync meta-codesync Bot changed the title Namespace lifted tensor constants per method in to_native (#23891) Namespace immutable constants per method in to_native (#23891) Oct 16, 2026
@JakeStevens @meta-codesync
Add CMake CI for native runtime, GGML, and Vulkan
7cebc6a 
Summary:
Add separate Linux jobs for `runtime`, `ggml`, and `vulkan`, running each native test suite once and building `run_classification` in the Vulkan job. Scope PR triggers to relevant build inputs.

Authored with AI assistance from Codex.

Differential Revision: D125695225
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 16, 2026
@JakeStevens @meta-codesync
Namespace immutable constants per method in to_native (pytorch#23891)
f807a6e 
Summary:
Pull Request resolved: pytorch#23891

Exported methods can reuse a constant name with different immutable tensors, causing `to_native` to reject the combined program. Namespace every immutable constant as `<method>/<key>`, including single-method programs, and update its graph references, including quantization metadata and nested graphs. Identical tensors still deduplicate when saving the `.ptn`.

Keys mutated by any method retain their original names throughout the program, preserving shared state for readers and writers. Reject differing values for shared mutable keys and collisions between final names.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428
@JakeStevens
JakeStevens force-pushed the export-D125294428 branch 2 times, most recently from f807a6e to fb49267 Compare October 16, 2026 18:27
JakeStevens added a commit to JakeStevens/executorch that referenced this pull request Oct 16, 2026
@JakeStevens @meta-codesync
Namespace immutable constants per method in to_native (pytorch#23891)
fb49267 
Summary:
Pull Request resolved: pytorch#23891

Exported methods can reuse a constant name with different immutable tensors, causing `to_native` to reject the combined program. Namespace every immutable constant as `<method>/<key>`, including single-method programs, and update its graph references, including quantization metadata and nested graphs. Identical tensors still deduplicate when saving the `.ptn`.

Keys mutated by any method retain their original names throughout the program, preserving shared state for readers and writers. Reject differing values for shared mutable keys and collisions between final names.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428
@JakeStevens @meta-codesync
Add CMake targets for HF LLM runners (pytorch#23899)
307f135 
Summary:
Pull Request resolved: pytorch#23899

Add `EXECUTORCH_BUILD_NATIVE_LLM_RUNNER` to build and install `run_hf_llm_vulkan` and `run_hf_llm_ggml` for the enabled engines. Share gflags setup with the vision runner, resolve tokenizers with regex lookahead support, and add `native_llm_runner_utils_test`. Document building and running the GGML CPU LLM runner.

Authored with assistance from Codex.

Differential Revision: D125479665
@JakeStevens @meta-codesync
Namespace immutable constants per method in to_native (pytorch#23891)
27fff9b 
Summary:
Pull Request resolved: pytorch#23891

Exported methods can reuse a constant name with different immutable tensors, causing `to_native` to reject the combined program. Namespace every immutable constant as `<method>/<key>`, including single-method programs, and update its graph references, including quantization metadata and nested graphs. Identical tensors still deduplicate when saving the `.ptn`.

Keys mutated by any method retain their original names throughout the program, preserving shared state for readers and writers. Reject differing values for shared mutable keys and collisions between final names.

Authored with assistance from OpenAI Codex.

Differential Revision: D125294428

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@digantdesai digantdesai digantdesai requested changes
@larryliu0820 larryliu0820 Awaiting requested review from larryliu0820 larryliu0820 is a code owner
@kirklandsign kirklandsign Awaiting requested review from kirklandsign
@JacobSzwejbka JacobSzwejbka Awaiting requested review from JacobSzwejbka JacobSzwejbka is a code owner
@SS-JIA SS-JIA Awaiting requested review from SS-JIA SS-JIA is a code owner

Requested changes must be addressed to merge this pull request.

Assignees

No one assigned

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. meta-exported

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@JakeStevens @digantdesai