Skip to content

Conversation

@pytorchbot
Copy link
Collaborator

This PR was created by the merge bot to help merge the original PR into the main branch.
ghstack PR number: #16399 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/78/base
ghstack PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/78/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/gasoonjia/77/orig
Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/78/orig
Differential Revision: D89821402
@diff-train-skip-merge

Pull Request resolved: #16398

This diff adds factory functions to create SlimTensor instances with allocated storage for better usage.

**Key components:**

1. **`factory/Empty.h`** - Factory functions for creating empty tensors:
   - `empty_strided(sizes, strides, dtype, device)` - Creates tensor with explicit sizes/strides
   - `empty(sizes, dtype, device)` - Creates contiguous tensor (auto-computes row-major strides)
   - `empty(initializer_list, dtype, device)` - Convenience overload for inline size specification
   - `empty_like(tensor)` - Creates empty tensor with same metadata as another tensor

2. **`core/Storage.h`** - Added `new_storage()` helper function:
   - Computes required storage bytes from sizes, strides, and dtype
   - Allocates and returns a new owning Storage instance
ghstack-source-id: 334622015
@exported-using-ghexport

Differential Revision: [D89820292](https://our.internmc.facebook.com/intern/diff/D89820292/)
…oat16)

Pull Request resolved: #16399

This diff adds support for all required scalar types in SlimTensor to support ExecuTorch aoti-driven backend usage: Int8 (Char), Int16 (Short), Int32 (Int), Int64 (Long), Bool, and BFloat16.

**Key changes:**

1. **`c10/core/ScalarType.h`** - Extended with all required types:
   - Added enum values matching PyTorch's c10::ScalarType for compatibility
   - Added type alias constants (kChar, kShort, kInt, kLong, kBool, kBFloat16)
   - Extended `elementSize()` to return correct sizes for all types
   - Extended `toString()` for all types
   - Fixed `isFloatingType()` to include BFloat16
   - Fixed `isIntegralType()` to properly handle all integral types and Bool
   - Added `isBoolType()` helper function
   - Imported BFloat16 from ExecuTorch's portable_type
ghstack-source-id: 334622008

Differential Revision: [D89821402](https://our.internmc.facebook.com/intern/diff/D89821402/)
@pytorch-bot
Copy link

pytorch-bot bot commented Jan 21, 2026

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16731

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

⏳ No Failures, 140 Pending

As of commit 5c8a9b1 with merge base 4f8dbde (image):
💚 Looks good so far! There are no failures yet. 💚

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

@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 Jan 21, 2026
Base automatically changed from gh/gasoonjia/77/orig to main January 21, 2026 06:10
@Gasoonjia Gasoonjia merged commit 9bddcfe into main Jan 21, 2026
154 of 160 checks passed
@Gasoonjia Gasoonjia deleted the gh/gasoonjia/78/orig branch January 21, 2026 06:13
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.

3 participants