Bump PyTorch pin to 2025-03-25

[swolchok]

Found release from https://hud2.pytorch.org/hud/pytorch/pytorch/nightly/0 .

[swolchok]

Stack from ghstack (oldest at bottom):

• Add optimized ELU implementation #9521
• -> Bump PyTorch pin to 2025-03-25 #9578

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/9578

• 📄 Preview Python docs built from this PR

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

❌ 2 New Failures

As of commit f78320d with merge base 1572381 ():

NEW FAILURES - The following jobs have failed:

• pull / unittest-arm / linux-job (gh)
  RuntimeError: Command docker exec -t 7d9365397adc8cc35c8aeac785f662a8602f1d60c88698c54971890b81376f29 /exec failed with exit code 1
• trunk / test-arm-reference-delegation / linux-job (gh)
  RuntimeError: Command docker exec -t f4b5843dfc3cc6ed207c1efe1be67ba9e50b55810d7edfa48a79e0449dd47a9f /exec failed with exit code 1

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

[mergennachin]

Can we do this after our branch cut on Wednesday 6pm?

The reason is that 3-12 is the RC candidate for PT2.7 - anything after that is in the PT2.8

Otherwise, we would have to go back and forth, instead of monotonically increasing the nightly.

[mergennachin]

Unfortunately, this is the downside of delaying the branch cut and diverging from pytorch timeline

[swolchok]

@mergennachin the branch is cut now, right?

[mergennachin]

Thank you for your patience, we can move forward now that its post branch cut

[mergennachin]

cc @metascroy

[swolchok]

looks like unittest-arm failure are new, CC @digantdesai

[swolchok]

resnapshotted the ARM test. we'll see if it builds.

[swolchok]

pinged @digantdesai for advice on how to proceed with ARM failures

[zingo]

We have looked into this and have some fixes in the Arm delegate in the works to handle this, hopefully it will be ready tomorrow.

[zingo]

Fix that hopefully handled the Arm backend problems merged to main e.g. #9834

[swolchok]

Fix that hopefully handled the Arm backend problems merged to main e.g. #9834

@zingo still one failure on unittest-arm:

FAILED backends/arm/test/models/test_torch_functions.py::test_torch_fns_BI[t] - RuntimeError: Unknown operator <EdgeOpOverload: quantized_decomposed.quantize_per_tensor.default>: schema = quantized_decomposed::quantize_per_tensor(Tensor input, float scale, int zero_point, int quant_min, int quant_max, ScalarType dtype) -> Tensor for TOSA : TOSA-0.80+BI