Remove stale ConvTranspose3d complex half parity xfail

[gplutop7]

Fixes:
#3030

This PR removes the stale inherited expected failure for ConvTranspose3d complex half CPU/GPU parity on XPU. The test now passes with the corresponding PyTorch change: pytorch/pytorch#179713

[Copilot]

Pull request overview

This PR updates the XPU test harness to stop inheriting a CUDA expectedFailure decoration for the nn.ConvTranspose3d CPU/GPU parity test, so that XPU no longer treats the test as an expected failure when it now passes.

Changes:

• Add ("nn.ConvTranspose3d", "test_cpu_gpu_parity") to the _cuda_xfail_xpu_pass allowlist so XPU doesn’t inherit CUDA’s expectedFailure for that test.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

Performance outliers, please check!

• 🔴 [-1, 80%), should be regression

Category	Model	Target vs. Baseline [Eager]	Target vs. Baseline [Inductor]
torchbench_bfloat16_training	densenet121	0.736544	0.654402
torchbench_bfloat16_training	mnasnet1_0	0.954740	0.749345
torchbench_bfloat16_training	resnet18	0.892120	0.779054

• 🟡 [80%, 90%), may be fluctuations

Category	Model	Target vs. Baseline [Eager]	Target vs. Baseline [Inductor]
torchbench_bfloat16_training	dcgan	0.850657	0.807232
torchbench_bfloat16_training	resnext50_32x4d	0.961751	0.839385
torchbench_bfloat16_training	mobilenet_v3_large	0.961968	0.858137