[ET-VK][ez] Improve insert_prepack_node pass to handle multiple uses of constant tensors #10426
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…of constant tensors ## Context Refer to #6352 for why the `insert_prepack_nodes` pass is needed. The current logic of the pass assumes that each constant tensor node has only one use. However, in reality, a constant tensor node may have multiple uses; some of which may require the insertion of a prepacking node and some which may not (since they may choose to apply some special packing in the operator implementation). Currently, if any uses of a constant tensor node handles its own prepacking, then prepacking nodes will not be inserted. This makes it so that a model will produce a type error during runtime when an operator receives a `TensorRef` but expects a `Tensor`. ## Changes Improve the logic of the pass to handle constant tensor nodes which have multiple uses. If any use does not handle its own prepacking, then a prepacking node will be inserted for those usages. Differential Revision: [D73592619](https://our.internmc.facebook.com/intern/diff/D73592619/) [ghstack-poisoned]
SS-JIA
added a commit
that referenced
this pull request
Apr 24, 2025
…of constant tensors ## Context Refer to #6352 for why the `insert_prepack_nodes` pass is needed. The current logic of the pass assumes that each constant tensor node has only one use. However, in reality, a constant tensor node may have multiple uses; some of which may require the insertion of a prepacking node and some which may not (since they may choose to apply some special packing in the operator implementation). Currently, if any uses of a constant tensor node handles its own prepacking, then prepacking nodes will not be inserted. This makes it so that a model will produce a type error during runtime when an operator receives a `TensorRef` but expects a `Tensor`. ## Changes Improve the logic of the pass to handle constant tensor nodes which have multiple uses. If any use does not handle its own prepacking, then a prepacking node will be inserted for those usages. Differential Revision: [D73592619](https://our.internmc.facebook.com/intern/diff/D73592619/) ghstack-source-id: 280054012 Pull Request resolved: #10426
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/10426
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New FailureAs of commit 73c5660 with merge base dfd3dbe ( NEW FAILURE - The following job has failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D73592619 |
trivedivivek
approved these changes
Apr 25, 2025
facebook-github-bot
merged commit 469c6f4
into
gh/SS-JIA/219/base
84 of 89 checks passed
facebook-github-bot
temporarily deployed
to
cherry-pick-bot
GitHub Actions
Inactive
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
Context
Refer to #6352 for why the
insert_prepack_nodespass is needed.The current logic of the pass assumes that each constant tensor node has only one use. However, in reality, a constant tensor node may have multiple uses; some of which may require the insertion of a prepacking node and some which may not (since they may choose to apply some special packing in the operator implementation).
Currently, if any uses of a constant tensor node handles its own prepacking, then prepacking nodes will not be inserted. This makes it so that a model will produce a type error during runtime when an operator receives a
TensorRefbut expects aTensor.Changes
Improve the logic of the pass to handle constant tensor nodes which have multiple uses. If any use does not handle its own prepacking, then a prepacking node will be inserted for those usages.
Differential Revision: D73592619