Skip to content

Add CI test for CoreML LoRA multimethod export#18354

Open
lucylq wants to merge 11 commits intogh/lucylq/148/headfrom
gh/lucylq/147/head
Open

Add CI test for CoreML LoRA multimethod export#18354
lucylq wants to merge 11 commits intogh/lucylq/148/headfrom
gh/lucylq/147/head

Conversation

@lucylq
Copy link
Copy Markdown
Contributor

@lucylq lucylq commented Mar 19, 2026

Tests three export configurations with file size validation:

  1. Base only (single method) - baseline size
  2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
  3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.

[ghstack-poisoned]
@lucylq
Copy link
Copy Markdown
Contributor Author

lucylq commented Mar 19, 2026

@pytorch-bot
Copy link
Copy Markdown

pytorch-bot Bot commented Mar 19, 2026

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18354

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

❌ 1 New Failure, 5 Cancelled Jobs, 3 Unrelated Failures

As of commit 1b2b985 with merge base e90d3c8 (image):

NEW FAILURE - The following job has failed:

CANCELLED JOBS - The following jobs were cancelled. Please retry:

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

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

lucylq added a commit that referenced this pull request Mar 19, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 7fea40d
ghstack-comment-id: 4094191365
Pull-Request: #18354
@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 Mar 19, 2026
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 19, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 79dd55a
ghstack-comment-id: 4094191365
Pull-Request: #18354
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 19, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 02d3ab3
ghstack-comment-id: 4094191365
Pull-Request: #18354
@lucylq lucylq changed the base branch from gh/lucylq/146/head to gh/lucylq/148/head March 19, 2026 23:49
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 20, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 736934f
ghstack-comment-id: 4094191365
Pull-Request: #18354
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 20, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: f09464e
ghstack-comment-id: 4094191365
Pull-Request: #18354
@lucylq lucylq changed the title Add CI test for CoreML LoRA multimethod export [wip] Add CI test for CoreML LoRA multimethod export Mar 20, 2026
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 20, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 30ed980
ghstack-comment-id: 4094191365
Pull-Request: #18354
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 20, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: ac64f26
ghstack-comment-id: 4094191365
Pull-Request: #18354
@lucylq lucylq changed the title [wip] Add CI test for CoreML LoRA multimethod export Add CI test for CoreML LoRA multimethod export Mar 20, 2026
@lucylq lucylq changed the title Add CI test for CoreML LoRA multimethod export [wip] Add CI test for CoreML LoRA multimethod export Mar 20, 2026
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 24, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 4d069b2
ghstack-comment-id: 4094191365
Pull-Request: #18354
@lucylq lucylq changed the title [wip] Add CI test for CoreML LoRA multimethod export Add CI test for CoreML LoRA multimethod export Mar 24, 2026
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 30, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 46ff4b0
ghstack-comment-id: 4094191365
Pull-Request: #18354
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Mar 31, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 95eedd9
ghstack-comment-id: 4094191365
Pull-Request: #18354
[ghstack-poisoned]
lucylq added a commit that referenced this pull request Apr 2, 2026
Tests three export configurations with file size validation:
1. Base only (single method) - baseline size
2. Base + LoRA adapter (multimethod) - small overhead from lora_a/lora_b
3. Base + LoRA + multifunction - same overhead (POSITIONAL sharing)

Uses stories110M with a synthetic zero-initialized LoRA adapter so
base and adapter outputs match. Inference tests run on macOS only.

Authored with Claude.


ghstack-source-id: 1b802dd
ghstack-comment-id: 4094191365
Pull-Request: #18354
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

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

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant