build(deps): Bump torch from 2.8.0 to 2.12.0 in /baselines/dev/baseline

[dependabot]

Bumps torch from 2.8.0 to 2.12.0.

Release notes

Sourced from torch's releases.

PyTorch 2.12.0 Release Notes

• Highlights
• Backwards Incompatible Changes
• Deprecations
• New Features
• Improvements
• Bug fixes
• Performance
• Documentation
• Developers
• Security

Highlights

For more details about these highlighted features, you can look at the release blogpost. Below are the full release notes for this release.

Backwards Incompatible Changes

Build Frontend

• Strengthened SVE compile checks in FindARM.cmake, which may reject previously accepted but incorrect SVE configurations (#176646)

  Source builds that enable SVE now validate the compiler configuration more strictly. If a build previously passed with an incomplete or mismatched SVE setup, it may now fail during CMake configuration instead of later in compilation. Update the compiler/toolchain flags so they accurately describe the target SVE support, or disable SVE for that build.

• Updated the minimum CUDA version required to build PyTorch from source to CUDA 12.6 (#178925)

  Building PyTorch from source with CUDA versions older than 12.6 is no longer supported. Users building custom binaries should install CUDA 12.6 or newer and make sure CUDA_HOME points to that installation.

  Version 2.11:

  CUDA_HOME=/usr/local/cuda-12.4 python setup.py develop

  Version 2.12:

  CUDA_HOME=/usr/local/cuda-12.6 python setup.py develop

• Enforced a C++20 minimum in CMake build files (#178662)

... (truncated)

Commits

• 0d62256 [release] Dockerfile: skip torchaudio install when CUDA_PATH=cu132 (#183346)
• 7661cd9 [MPS] Fix SDPA wrong output for permuted q/k/v with B > 1 (#181886)
• 9da6087 Fix stale PYTORCH_RELEASES_CODE_CC dict (fixes #182250) (#182369)
• e4c37cc Avoid raw stream name collisions in Inductor (#182178)
• 822d047 [MPS] Fix bool mask handling in 1-pass SDPA decode kernel (#182285) (#182311)
• 5c5e523 Add enable_gqa parameter to SDPA MPS meta registration (#181550)
• eece52e [AOTI] Add BC-safe c_shim v2 for _scaled_dot_product_attention_math_for_mps e...
• b39330b [Inductor] Call latest c_shim version for versioned fallback ops (#181548)
• 06f10d0 Revert "[FSDP2] add fqn to communication ops" (#182100) (#182157)
• 449e339 Revert "[Inductor] Improve materialization heuristic for a chain of computaio...
• Additional commits viewable in compare view

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 558b8aa9fe

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Keep torchvision pinned compatibly with torch

With this exact pin, the baseline template now asks resolvers to install torch==2.12.0 while still pinning torchvision==0.23.0; that torchvision release is paired with and declares an exact dependency on torch==2.8.0. In a normal pip/uv install of baselines/dev/baseline, dependency resolution becomes unsatisfiable instead of producing a working template, so the torchvision pin needs to be bumped to the release matching torch 2.12.0 (or both pins kept together).

Useful? React with 👍 / 👎.