Introduce data schema to store raw tensors #6540

lucylq · 2024-10-28T22:39:32Z

Stack from ghstack (oldest at bottom):

-> Introduce data schema to store raw tensors #6540

Differential Revision: D65156641

[ghstack-poisoned]

pytorch-bot · 2024-10-28T22:39:35Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6540

📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

[DomainsOnly] Jobs fail with GLIBC version not found

✅ You can merge normally! (1 Unrelated Failure)

As of commit 70c9f7f with merge base 3813f42 ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

Propose to merge ghstack orig PRs to main / Try to create a PR with ghstack /orig branch (gh) (detected as infra flaky with no runner)

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

ghstack-source-id: da15b0f Pull Request resolved: #6540

lucylq · 2024-10-28T22:40:39Z

See: #6492

[ghstack-poisoned]

ghstack-source-id: 99f32da Pull Request resolved: #6540

schema/data.fbs

+  // Offset in scalar_type elements (e.g., multiples of 4 bytes for an int
+  // scalar type) from the beginning of the tensor buffer to the beginning of
+  // the actual data. Currently, the runtime only supports a value of zero.
+  storage_offset:int;


lucylq · 2024-10-28T23:09:18Z

schema/data.fbs

+  scalar_type:ScalarType;
+
+  // Size of each dimension.
+  dim_sizes:[int];


equivalent of sizes in program.fbs, using dim_sizes to differentiate from size of the tensor below.

lucylq · 2024-10-29T17:56:42Z

@lucylq has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

ghstack-source-id: 14770e7 Pull Request resolved: #6540

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

ghstack-source-id: 5f83f21 Pull Request resolved: #6540

lucylq · 2024-10-30T23:19:53Z

@lucylq has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

facebook-github-bot · 2024-10-30T23:37:33Z

This pull request was exported from Phabricator. Differential Revision: D65156641

Pull Request resolved: #6540 @imported-using-ghimport Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641/) ghstack-source-id: 251037652

JacobSzwejbka

I was thinking maybe we should try to reuse more fields from the core schema but I guess the goal of this is different enough that we dont have to. Are we 100% sure we even need sizes and dim order. I think the set of fully qualified names provides a good enough compatibility guarantee (seems unlikely 2 models would have the exact same key set but different values)

JacobSzwejbka · 2024-11-05T19:23:24Z

exir/schema_data.py

+class TensorMetadata:
+    fully_qualified_name: str
+    scalar_type: ScalarType
+    dim_sizes: List[int]


nit: just sizes not dim_sizes

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

Pull Request resolved: #6540 @imported-using-ghimport Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641/) ghstack-source-id: 253116779

facebook-github-bot · 2024-11-12T16:19:17Z

This pull request was exported from Phabricator. Differential Revision: D65156641

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

Pull Request resolved: #6540 @imported-using-ghimport Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641/) ghstack-source-id: 253187679

facebook-github-bot · 2024-11-12T21:19:04Z

This pull request was exported from Phabricator. Differential Revision: D65156641

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

Pull Request resolved: #6540 @imported-using-ghimport Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641/) ghstack-source-id: 253188153

facebook-github-bot · 2024-11-12T21:21:11Z

This pull request was exported from Phabricator. Differential Revision: D65156641

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

Pull Request resolved: #6540 @imported-using-ghimport Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641/) ghstack-source-id: 253910001

facebook-github-bot · 2024-11-16T03:25:19Z

This pull request was exported from Phabricator. Differential Revision: D65156641

[executorch][serialization] Introduce data schema to store raw tensors Pull Request resolved: #6540 @imported-using-ghimport Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641/) ghstack-source-id: 253910001

Introduce data schema to store raw tensors

f24016d

[ghstack-poisoned]

lucylq added a commit that referenced this pull request Oct 28, 2024

Introduce data schema to store raw tensors

78d081b

ghstack-source-id: da15b0f Pull Request resolved: #6540

facebook-github-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 28, 2024

lucylq requested review from JacobSzwejbka and dbort October 28, 2024 22:39

Update on "Introduce data schema to store raw tensors"

229527e

[ghstack-poisoned]

lucylq added a commit that referenced this pull request Oct 28, 2024

Introduce data schema to store raw tensors

514cc22

ghstack-source-id: 99f32da Pull Request resolved: #6540

lucylq commented Oct 28, 2024

View reviewed changes

lucylq requested a review from iseeyuan October 28, 2024 23:14

Update on "Introduce data schema to store raw tensors"

cec36f6

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

lucylq added a commit that referenced this pull request Oct 29, 2024

Introduce data schema to store raw tensors

de44664

ghstack-source-id: 14770e7 Pull Request resolved: #6540

Update on "Introduce data schema to store raw tensors"

859cf52

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

lucylq added a commit that referenced this pull request Oct 30, 2024

Introduce data schema to store raw tensors

3c2c7ab

ghstack-source-id: 5f83f21 Pull Request resolved: #6540

Update on "Introduce data schema to store raw tensors"

c8d6489

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

facebook-github-bot added the fb-exported label Oct 30, 2024

JacobSzwejbka approved these changes Nov 5, 2024

View reviewed changes

Update on "Introduce data schema to store raw tensors"

d7b7162

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

This was referenced Nov 12, 2024

[executorch][serialization] Rename scalar_type.fbs --> common.fbs #6678

Closed

[executorch][serialization] Move DataSegment into shared common.fbs #6679

Closed

Update on "Introduce data schema to store raw tensors"

9043914

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

Update on "Introduce data schema to store raw tensors"

c829c57

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

dbort approved these changes Nov 13, 2024

View reviewed changes

Update on "Introduce data schema to store raw tensors"

70c9f7f

Differential Revision: [D65156641](https://our.internmc.facebook.com/intern/diff/D65156641) [ghstack-poisoned]

facebook-github-bot merged commit 2548aeb into gh/lucylq/12/base Nov 16, 2024
41 checks passed

facebook-github-bot deleted the gh/lucylq/12/head branch November 16, 2024 04:40

facebook-github-bot had a problem deploying to cherry-pick-bot November 16, 2024 04:40 — with GitHub Actions Failure

lucylq mentioned this pull request Nov 16, 2024

Introduce data schema to store raw tensors #6919

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Introduce data schema to store raw tensors #6540

Introduce data schema to store raw tensors #6540

Uh oh!

lucylq commented Oct 28, 2024 •

edited

Loading

Uh oh!

pytorch-bot bot commented Oct 28, 2024 •

edited

Loading

Uh oh!

lucylq commented Oct 28, 2024

Uh oh!

This comment was marked as resolved.

Uh oh!

lucylq Oct 28, 2024 •

edited

Loading

Uh oh!

lucylq commented Oct 29, 2024

Uh oh!

lucylq commented Oct 30, 2024

Uh oh!

facebook-github-bot commented Oct 30, 2024

Uh oh!

JacobSzwejbka left a comment

Uh oh!

JacobSzwejbka Nov 5, 2024 •

edited

Loading

Uh oh!

facebook-github-bot commented Nov 12, 2024

Uh oh!

facebook-github-bot commented Nov 12, 2024

Uh oh!

facebook-github-bot commented Nov 12, 2024

Uh oh!

facebook-github-bot commented Nov 16, 2024

Uh oh!

Uh oh!

Uh oh!

Introduce data schema to store raw tensors #6540

Introduce data schema to store raw tensors #6540

Uh oh!

Conversation

lucylq commented Oct 28, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Oct 28, 2024 • 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/6540

❗ 1 Active SEVs

✅ You can merge normally! (1 Unrelated Failure)

Uh oh!

lucylq commented Oct 28, 2024

Uh oh!

This comment was marked as resolved.

Uh oh!

lucylq Oct 28, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lucylq commented Oct 29, 2024

Uh oh!

lucylq commented Oct 30, 2024

Uh oh!

facebook-github-bot commented Oct 30, 2024

Uh oh!

JacobSzwejbka left a comment

Choose a reason for hiding this comment

Uh oh!

JacobSzwejbka Nov 5, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

facebook-github-bot commented Nov 12, 2024

Uh oh!

facebook-github-bot commented Nov 12, 2024

Uh oh!

facebook-github-bot commented Nov 12, 2024

Uh oh!

facebook-github-bot commented Nov 16, 2024

Uh oh!

Uh oh!

Uh oh!

lucylq commented Oct 28, 2024 •

edited

Loading

pytorch-bot bot commented Oct 28, 2024 •

edited

Loading

lucylq Oct 28, 2024 •

edited

Loading

JacobSzwejbka Nov 5, 2024 •

edited

Loading