Support multimethod in export_llama_lib by lucylq · Pull Request #17231 · pytorch/executorch

lucylq · 2026-02-05T02:22:16Z

Stack from ghstack (oldest at bottom):

TODO: add CI test.

Differential Revision: D92315602

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

pytorch-bot · 2026-02-05T02:22:20Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17231

📄 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, 4 Unrelated Failures

As of commit 4aa2372 with merge base aa2f683 ():

NEW FAILURE - The following job has failed:

trunk / test-static-llama-qnn-eval-linux (baseline, 62) / linux-job (gh)
RuntimeError: Command docker exec -t a1da39a4ee0b07f02f8e5e88b134a9496060712fde5c3f6e59f755d2a73e8d83 /exec failed with exit code 1

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

trunk / test-models-windows (emformer_transcribe, xnnpack-q8) / windows-job (gh) (detected as infra flaky with no log or failing log classifier)
trunk / test-models-windows (mv3, xnnpack-q8) / windows-job (gh) (detected as infra flaky with no log or failing log classifier)
trunk / test-models-windows (resnet50, portable) / windows-job (gh) (detected as infra flaky with no log or failing log classifier)
trunk / test-models-windows (resnet50, xnnpack-q8) / windows-job (gh) (detected as infra flaky with no log or failing log classifier)

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

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) ghstack-source-id: 338339885 Pull Request resolved: #17231

github-actions · 2026-02-05T02:22:53Z

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.

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 TODO: add CI test. Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration ghstack-source-id: 338339885 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

Copilot

Pull request overview

This pull request adds multimethod export support to the Llama export library, enabling the export of multiple methods (base model and LoRA variants) into a single .pte file. This is part of a stack of PRs (#17228-#17231) that collectively add multimethod and LoRA support to the ExecuTorch runtime.

Changes:

Added _export_llama_multimethod() function to handle exporting multiple methods to a single .pte file
Added helper functions _get_xnnpack_partitioners() and _get_output_filename() to support multimethod export
Added validation logic to ensure multimethod export only works with XNNPACK backend or portable ops
Added configuration file for Qwen3 multimethod export with LoRA
Added comprehensive test script for validating multimethod export functionality
Updated build dependencies to include required tokenizer and weight conversion modules

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
`examples/models/llama/export_llama_lib.py`	Core implementation of multimethod export logic, including validation, helper functions, and the main export function
`examples/models/qwen3/config/qwen3_multimethod.yaml`	Configuration file demonstrating multimethod export with LoRA and base methods using environment variable interpolation
`examples/models/llama/BUCK`	Added `convert_weights.py` to export library sources for proper dependency resolution
`examples/models/llama/runner/targets.bzl`	Added `regex_lookahead` tokenizer dependency for enhanced tokenizer support
`.ci/scripts/test_lora_multimethod.sh`	Comprehensive test script validating both LoRA and base method execution with expected output verification

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration ghstack-source-id: 339710072 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

kimishpatel

Review automatically exported from Phabricator review in Meta.

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 340935057 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341144063 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341145198 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341145198 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341145198 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341145198 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341235800 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Pull Request resolved: #17231 Note: multimethod export is currently limited to: - xnnpack or portable lib - only lora (does not support arbitrary nn.Modules in each method) - if quant is enabled, lora models must share quant schemes at source transformation time - no pt2e quant, as each model could have slightly different results after calibration Changes: 1. Add MultimethodLoraConfig to yaml 2. Deepcopy yaml config. Move each lora_config into base.lora_config. 3. Create and export the model 4. Repeat 2,3 for each method. 5. Pass a dict of method_name: ep to `to_edge_transform_and_lower` ghstack-source-id: 341235800 @exported-using-ghexport Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/)

@lucylq

This PR was created by the merge bot to help merge the original PR into the main branch. ghstack PR number: #17231 by @lucylq ^ Please use this as the source of truth for the PR details, comments, and reviews ghstack PR base: https://github.com/pytorch/executorch/tree/gh/lucylq/134/base ghstack PR head: https://github.com/pytorch/executorch/tree/gh/lucylq/134/head Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/lucylq/133/orig Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/lucylq/134/orig Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) @diff-train-skip-merge --------- Co-authored-by: Github Executorch <github_executorch@arm.com> Co-authored-by: lucylq <lfq@meta.com>

Support multimethod in export_llama_lib

854e3b5

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

This was referenced Feb 5, 2026

Support multimethod in runner #17228

Merged

Introduce LoraConfig #17229

Merged

Introduce MultimethodLoraConfig #17230

Merged

lucylq pushed a commit that referenced this pull request Feb 5, 2026

Support multimethod in export_llama_lib

bddf6dc

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) ghstack-source-id: 338339885 Pull Request resolved: #17231

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 Feb 5, 2026

meta-codesync Bot added fb-exported meta-exported labels Feb 5, 2026

lucylq mentioned this pull request Feb 5, 2026

Support multimethod in export_llama #17010

Closed

lucylq changed the title ~~Support multimethod in export_llama_lib~~ [wip] Support multimethod in export_llama_lib Feb 5, 2026

Update on "[wip] Support multimethod in export_llama_lib"

1de0598

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

lucylq requested a review from Copilot February 5, 2026 19:01

Copilot started reviewing on behalf of lucylq February 5, 2026 19:02 View session

Copilot AI reviewed Feb 5, 2026

View reviewed changes

Comment thread examples/models/llama/export_llama_lib.py Outdated

Comment thread examples/models/llama/export_llama_lib.py Outdated

Comment thread examples/models/llama/export_llama_lib.py

Comment thread .ci/scripts/test_lora_multimethod.sh

Update on "[wip] Support multimethod in export_llama_lib"

f0522dc

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

lucylq changed the title ~~[wip] Support multimethod in export_llama_lib~~ Support multimethod in export_llama_lib Feb 10, 2026

kimishpatel approved these changes Feb 13, 2026

View reviewed changes

Update on "Support multimethod in export_llama_lib"

5853a6b

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

lucylq requested review from larryliu0820 and mergennachin as code owners February 13, 2026 06:40

Update on "Support multimethod in export_llama_lib"

4b51baf

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Update on "Support multimethod in export_llama_lib"

11e4f7a

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Update on "Support multimethod in export_llama_lib"

542fa80

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Update on "Support multimethod in export_llama_lib"

c2acf0c

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Update on "Support multimethod in export_llama_lib"

1908b30

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Update on "Support multimethod in export_llama_lib"

320e3aa

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

Update on "Support multimethod in export_llama_lib"

4aa2372

TODO: add CI test. Differential Revision: [D92315602](https://our.internmc.facebook.com/intern/diff/D92315602/) [ghstack-poisoned]

meta-codesync Bot merged commit 943e869 into gh/lucylq/134/base Feb 13, 2026
334 of 341 checks passed

meta-codesync Bot deleted the gh/lucylq/134/head branch February 13, 2026 23:59

meta-codesync Bot temporarily deployed to cherry-pick-bot February 13, 2026 23:59 Inactive

pytorchbot mentioned this pull request Feb 13, 2026

Support multimethod in export_llama_lib #17462

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support multimethod in export_llama_lib#17231

Support multimethod in export_llama_lib#17231
meta-codesync[bot] merged 11 commits intogh/lucylq/134/basefrom
gh/lucylq/134/head

lucylq commented Feb 5, 2026 •

edited

Loading

Uh oh!

pytorch-bot Bot commented Feb 5, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Feb 5, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

kimishpatel left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

lucylq commented Feb 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot Bot commented Feb 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17231

❌ 1 New Failure, 4 Unrelated Failures

Uh oh!

github-actions Bot commented Feb 5, 2026

This PR needs a release notes: label

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

kimishpatel left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

lucylq commented Feb 5, 2026 •

edited

Loading

pytorch-bot Bot commented Feb 5, 2026 •

edited

Loading

This PR needs a `release notes:` label