Issue Station
Sign In
Tools hereToolbox

pytorch - ✅(Solved) Fix [RFC] Drop macOS < 14.0 support for PyTorch 2.12 [2 pull requests, 3 comments, 4 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
pytorch/pytorch#177303•Fetched 2026-04-08 00:42:15
View on GitHub
Comments
3
Participants
4
Timeline
15
Reactions
2
Author
atalmanatalman
Participants
atalmanatalmanemmapaul357-afkemmapaul357-afkuwu-420uwu-420yangw-devyangw-dev
Timeline (top)
cross-referenced ×4commented ×3referenced ×3added_to_project_v2 ×1

Fix Action

Fix / Workaround

  • macOS 11 (Big Sur), 12 (Monterey), and 13 (Ventura) are all end-of-life and no longer receive security updates from Apple.
    • macOS 14 (Sonoma) is the oldest version still receiving active security patches.
    • Apple Silicon (arm64) Macs all support macOS 14.0, so this does not reduce hardware coverage for arm64 wheels.
    • This aligns our support matrix with Apple's own support lifecycle and reduces the maintenance burden of targeting older deployment targets.
    • Key dependencies are already moving to higher minimum macOS versions, making it increasingly difficult to maintain older compatibility without compromising the build.

PR fix notes

PR #7916: Bump MACOSX_DEPLOYMENT_TARGET to 14.0

Description (problem / solution / changelog)

Summary

https://github.com/pytorch/pytorch/issues/177303

Files changed

  • tools/pkg-helpers/pytorch_pkg_helpers/macos.py
  • tools/pkg-helpers/tests/test_macos.py

Test plan

  • python -m pytest tools/pkg-helpers/tests/test_macos.py passes

Co-authored-by: Claude [email protected]

Changed files

  • tools/pkg-helpers/pytorch_pkg_helpers/macos.py (modified, +1/-5)
  • tools/pkg-helpers/tests/test_macos.py (modified, +7/-7)

PR #179083: [CI] Bump MACOSX_DEPLOYMENT_TARGET to 14.0

Description (problem / solution / changelog)

Summary

https://github.com/pytorch/pytorch/issues/177303

  • Update MACOSX_DEPLOYMENT_TARGET from 11.0/11.1 to 14.0 across all macOS CI build scripts
  • Update wheel platform name from macosx_11_0_arm64 to macosx_14_0_arm64
  • Companion PR in test-infra: https://github.com/pytorch/test-infra/pull/7916

Files changed

  • .ci/pytorch/macos-common.sh — 11.1 → 14.0
  • .ci/pytorch/macos-build.sh — 11.0 → 14.0, plat-name updated
  • .ci/wheel/build_wheel.sh — mac_version and MACOSX_DEPLOYMENT_TARGET updated

Changed files

  • .ci/pytorch/macos-build.sh (modified, +1/-1)
  • .ci/pytorch/macos-common.sh (modified, +1/-1)
  • .ci/pytorch/smoke_test/check_wheel_tags.py (modified, +8/-2)
  • .ci/wheel/build_wheel.sh (modified, +3/-2)
RAW_BUFFERClick to expand / collapse

Starting with PyTorch 2.12, we plan to set the minimum supported macOS version to 14.0 (Sonoma), up from the current 11.0 (Big Sur).

Reference issue: https://github.com/pytorch/pytorch/issues/177140

Background:

As discovered in #177140, our macOS arm64 wheels have carried dependencies (e.g., libomp.dylib) that already require macOS 14.0 at the binary level, despite the wheel metadata advertising macOS 11.0 compatibility. While we fixed this mismatch for 2.11 by rebuilding libomp with the correct deployment target, maintaining compatibility with macOS versions that Apple no longer supports adds ongoing build complexity and risk of similar mismatches in the future.

Rationale:

  • macOS 11 (Big Sur), 12 (Monterey), and 13 (Ventura) are all end-of-life and no longer receive security updates from Apple.
  • macOS 14 (Sonoma) is the oldest version still receiving active security patches.
  • Apple Silicon (arm64) Macs all support macOS 14.0, so this does not reduce hardware coverage for arm64 wheels.
  • This aligns our support matrix with Apple's own support lifecycle and reduces the maintenance burden of targeting older deployment targets.
  • Key dependencies are already moving to higher minimum macOS versions, making it increasingly difficult to maintain older compatibility without compromising the build.

What changes:

  • MACOSX_DEPLOYMENT_TARGET will be set to 14.0 for all macOS builds.
  • Wheels will be tagged macosx_14_0_arm64 instead of macosx_11_0_arm64.
  • Users on macOS < 14.0 will need to either upgrade their OS or stay on PyTorch 2.11.

extent analysis

Fix Plan

To address the issue of PyTorch dropping support for macOS versions below 14.0, follow these steps:

  • Upgrade macOS: If possible, upgrade your macOS version to 14.0 (Sonoma) or higher to ensure compatibility with the latest PyTorch versions.
  • Stay on PyTorch 2.11: If upgrading macOS is not feasible, you can stay on PyTorch 2.11, which still supports older macOS versions.
  • Modify Wheel Tags: Update your wheel tags to macosx_14_0_arm64 to reflect the new minimum supported macOS version.

Code Changes

No specific code changes are required, but you may need to update your setup.py or setup.cfg files to reflect the new MACOSX_DEPLOYMENT_TARGET version:

import os

# Set the MACOSX_DEPLOYMENT_TARGET environment variable
os.environ['MACOSX_DEPLOYMENT_TARGET'] = '14.0'

Alternatively, you can modify your CMakeLists.txt file to set the CMAKE_OSX_DEPLOYMENT_TARGET variable:

set(CMAKE_OSX_DEPLOYMENT_TARGET "14.0")

Verification

To verify that the fix worked, check that your PyTorch installation is compatible with your macOS version by running:

import torch
print(torch.__version__)

If you're using the correct version of PyTorch, this should print the version number without any errors.

Extra Tips

  • Always check the PyTorch documentation for the latest compatibility information and supported macOS versions.
  • Consider using a virtual environment or containerization to manage different PyTorch versions and dependencies.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#ssr#installation#tensor shape#autograd error#callback error#memory management#API rate limit#retriever error#indexing error

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING