[slimtensor] Add item() methods with CUDA and CPU support for scalar value extraction #16840
+2,382
−20
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
…emory Pull Request resolved: #16442 Add from_blob() factory function that creates SlimTensor from external memory without taking ownership. This enables creating tensors that wrap user-provided buffers for use in AOTI model execution. Changes: - Create factory/FromBlob.h with from_blob() overloads - Support sizes/strides as ArrayRef or initializer_list - Support custom storage offset - Add from_blob target to factory/targets.bzl ghstack-source-id: 335280566 @exported-using-ghexport Differential Revision: [D89950482](https://our.internmc.facebook.com/intern/diff/D89950482/)
Pull Request resolved: #16443 Add as_strided() and as_strided_() view operations that allow creating views of a tensor with different sizes, strides, and storage offset. Views share the same underlying storage. Changes: - Create c10/core/WrapDimMinimal.h for dimension wrapping - Create core/SlimTensorView-incl.h with as_strided implementations - Add as_strided() and as_strided_() methods to SlimTensor - Add set_sizes_and_strides() overload that takes storage_offset - Update build targets to include new headers ghstack-source-id: 335280569 @exported-using-ghexport Differential Revision: [D89951036](https://our.internmc.facebook.com/intern/diff/D89951036/)
Pull Request resolved: #16444 Add permute(), reshape(), and clone_contiguous() operations for tensor shape manipulation. These operations enable dimension reordering and reshaping with automatic fallback to contiguous copy when views are not possible. ghstack-source-id: 335280559 @exported-using-ghexport Differential Revision: [D89952083](https://our.internmc.facebook.com/intern/diff/D89952083/)
…value extraction Pull Request resolved: #16445 Add two `item()` methods to SlimTensor for extracting scalar values from single-element tensors: 1. **`item<T>()`** - Templated version requiring explicit type specification 2. **`item()`** - Non-templated version that auto-detects dtype and returns double Both methods: - Support all 7 slim dtypes (Float, BFloat16, Long, Int, Short, Char, Bool) - Handle CUDA tensors by automatically copying data to CPU - Validate that tensor has exactly 1 element ghstack-source-id: 335280577 @exported-using-ghexport Differential Revision: [D90034340](https://our.internmc.facebook.com/intern/diff/D90034340/)
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16840
Note: Links to docs will display an error until the docs builds have been completed. ❌ 3 New Failures, 2 Unrelated FailuresAs of commit bb7a6a1 with merge base c7c4073 ( NEW FAILURES - The following jobs have failed:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
BROKEN TRUNK - The following job failed but was 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. |
meta-cla
bot
added
the
CLA Signed
Gasoonjia
approved these changes
Jan 23, 2026
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.
This PR was created by the merge bot to help merge the original PR into the main branch.
ghstack PR number: #16445 by @Gasoonjia
^ 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/gasoonjia/87/base
ghstack PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/87/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/gasoonjia/86/orig
Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/87/orig
Differential Revision: D90034340
@diff-train-skip-merge