Add MLX hardtanh op handler

[atharvjairath]

Fixes #18921

Adds MLX delegate support for aten.hardtanh.default by lowering it to the existing ClipNode path with the operator's min_val and max_val bounds. This enables bounded activation models, including ReLU6-style hardtanh usage, to stay delegated to MLX instead of failing as an unsupported op.

This also adds focused MLX op tests for:

• default hardtanh bounds [-1.0, 1.0]
• ReLU6 bounds [0.0, 6.0]
• symmetric custom bounds [-2.0, 2.0]
• asymmetric custom bounds [-0.25, 0.75]

Test plan:

lintrunner backends/mlx/ops.py backends/mlx/test/test_ops.py

ok No lint issues.

CPLUS_INCLUDE_PATH=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/c++/v1 python -m executorch.backends.mlx.test.run_all_tests --rebuild hardtanh

Rebuilding op_test_runner in /Users/atharvjairath/Desktop/timepass/executorch/cmake-out...
Build succeeded.

Running test: hardtanh_min-1.0_max1.0_2x3x4
✓ MLX delegation verified
C++ binary output: OK
✓ PASSED: All outputs match

Running test: hardtanh_min0.0_max6.0_4x8
✓ MLX delegation verified
C++ binary output: OK
✓ PASSED: All outputs match

Running test: hardtanh_min-2.0_max2.0_10
✓ MLX delegation verified
C++ binary output: OK
✓ PASSED: All outputs match

Running test: hardtanh_min-0.25_max0.75_2x8x16
✓ MLX delegation verified
C++ binary output: OK
✓ PASSED: All outputs match

TEST SUMMARY
Passed: 4
Failed: 0

This follows up on #18986 by adding custom min/max bound coverage and including the requested local test output.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19776

• 📄 Preview Python docs built from this PR

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

❗ 2 Active SEVs

There are 2 currently active SEVs. If your PR is affected, please view them below:

• /easycla is not responding
• LF Fleet unable to register self-hosted runners with pytorch/pytorch and pytorch/test-infra

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

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: atharvjairath / name: atharvjairath (19fa148)

[atharvjairath]

@pytorchbot label "release notes: apple"

[atharvjairath]

cc: @nil-is-all

[atharvjairath]

Friendly bump, PR is ready for review. Added MLX support for aten.hardtanh.default via the existing ClipNode path, with tests for default/custom bounds and ReLU6 usage.

Would appreciate a review when possible. cc @metascroy