Arm backend: Add support for ELU.default operator

[agrima1304]

Decomposes elu into other operators/ lookup table for FP and INT cases.

cc @digantdesai @freddan80 @per @zingo @oscarandersson8218

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13683

• 📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• ROCm MI2xx CI/CD workflows failing due to : download from https://api.github.com/repos/pytorch/pytorch timed out.

❌ 1 New Failure, 1 Cancelled Job, 9 Unrelated Failures

As of commit c3e6386 with merge base deaf37f ():

NEW FAILURE - The following job has failed:

• Build documentation / build (buck2) / Build doc (gh)
  At least one of the pre-conditions you specified did not hold

CANCELLED JOB - The following job was cancelled. Please retry:

• trunk / test-huggingface-transformers-coreml (qwen3-1.7b|xnnpack|--quantize) / macos-job (gh)
  ##[error]The operation was canceled.

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• trunk / test-llama-runner-mac (fp32, coreml) / macos-job (gh) (similar failure)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 2

BROKEN TRUNK - The following jobs failed but was present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• trunk / test-arm-backend (test_models_ethos-u55) / linux-job (gh) (trunk failure)
  AttributeError: 'bool' object has no attribute 'lower'
• trunk / test-arm-backend (test_models_ethos-u85) / linux-job (gh) (trunk failure)
  AttributeError: 'bool' object has no attribute 'lower'
• trunk / test-arm-backend (test_models_tosa) / linux-job (gh) (trunk failure)
  AttributeError: 'bool' object has no attribute 'lower'
• trunk / test-llama-runner-mac (fp32, mps) / macos-job (gh) (trunk failure)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 2
• trunk / test-llama-runner-mac (fp32, xnnpack+custom+quantize_kv) / macos-job (gh) (trunk failure)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 2
• trunk / test-llama-torchao-lowbit / macos-job (gh) (trunk failure)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 2
• trunk / test-mcu-models (cmake) / linux-job (gh) (trunk failure)
  AttributeError: 'bool' object has no attribute 'lower'
• trunk / test-models-macos-cpu (mv3, portable) / macos-job (gh) (trunk failure)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 139

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

[agrima1304]

@pytorchbot label "partner: arm"

[agrima1304]

@pytorchbot label "release notes: arm"

[agrima1304]

@pytorchbot label "ciflow/trunk"

[pytorch-bot]

To add these label(s) (ciflow/trunk) to the PR, please first approve the workflows that are awaiting approval (scroll to the bottom of this page).

This helps ensure we don't trigger CI on this PR until it is actually authorized to do so. Please ping one of the reviewers if you do not have access to approve and run workflows.

[shoumikhin]

@agrima1304 please check is the following Pyre type checks are resolved:

    [executorch/backends/nxp/aten_passes/neutron_aten_pass_manager.py:28:23] Incompatible variable type [9]: passes is declared to have type `List[Type[typing.Callable[[GraphModule], PassResult]]]` but is used as type `None`.
    [executorch/backends/nxp/aten_passes/neutron_aten_pass_manager.py:29:8] Incompatible variable type [9]: passes is declared to have type `List[Type[typing.Callable[[GraphModule], PassResult]]]` but is used as type `Union[List[Type[typing.Callable[[GraphModule], PassResult]]], List[Union[FuseBatchNormWithConvPass, FuseBatchNormWithLinearPass, SplitGroupConvolution]]]`.
    [executorch/backends/nxp/aten_passes/neutron_aten_pass_manager.py:35:25] Incompatible parameter type [6]: In call `PassManager.__init__`, for 1st positional argument, expected `Union[None, List[typing.Callable[[GraphModule], Optional[PassResult]]], List[List[typing.Callable[[GraphModule], Optional[PassResult]]]]]` but got `List[Type[typing.Callable[[GraphModule], PassResult]]]`.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:45:4] Uninitialized attribute [13]: Attribute `module` is declared in class `SplitGroupConvolution` to have type `torch.fx.graph_module.GraphModule` but is never initialized.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:174:47] Incompatible parameter type [6]: In call `torch.fx.graph.Graph.inserting_after`, for 1st positional argument, expected `Optional[node.Node]` but got `_C.Node`.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:182:8] Incompatible return type [7]: Expected `_C.Node` but got `node.Node`.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:184:4] Inconsistent override [14]: `executorch.backends.nxp.aten_passes.split_group_convolution.SplitGroupConvolution.call` overrides method defined in `PassBase` inconsistently. Could not find parameter `graph_module` in overriding signature.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:240:52] Incompatible parameter type [6]: In call `SplitGroupConvolution._create_parameter_node_for_data`, for 3rd positional argument, expected `_C.Node` but got `node.Node`.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:246:50] Incompatible parameter type [6]: In call `SplitGroupConvolution._create_parameter_node_for_data`, for 3rd positional argument, expected `_C.Node` but got `node.Node`.
    [executorch/backends/nxp/aten_passes/split_group_convolution.py:254:20] Incompatible parameter type [6]: In call `SplitGroupConvolution._get_topologically_last_node`, for 1st positional argument, expected `List[node.Node]` but got `List[Union[_C.Node, node.Node]]`.
    [executorch/backends/nxp/tests/models.py:87:12] Incompatible parameter type [6]: In call `torch.nn.modules.conv.Conv3d.__init__`, for argument `kernel_size`, expected `Union[Tuple[int, int, int], int]` but got `Union[Tuple[int, int], int]`.
    [executorch/backends/nxp/tests/models.py:88:12] Incompatible parameter type [6]: In call `torch.nn.modules.conv.Conv3d.__init__`, for argument `stride`, expected `Union[Tuple[int, int, int], int]` but got `Union[Tuple[int, int], int]`.
    [executorch/backends/nxp/tests/models.py:89:12] Incompatible parameter type [6]: In call `torch.nn.modules.conv.Conv3d.__init__`, for argument `padding`, expected `Union[Tuple[int, int, int], int, str]` but got `Union[Collection[int], int, str]`.
    [executorch/backends/nxp/tests/models.py:90:12] Incompatible parameter type [6]: In call `torch.nn.modules.conv.Conv3d.__init__`, for argument `dilation`, expected `Union[Tuple[int, int, int], int]` but got `Union[Tuple[int, int], int]`.

[zingo]

Hi @shoumikhin we look into it when back to work but first just double checking did you post this in the right PR?
The files you listed are is in the NXP driver, and this PR only changed Arm backend code, that switch checking to mypy and not pyre. But maybe the NXP code is using/calling passes/code from the arm backend?

[shoumikhin]

@zingo indeed, they are related to a different PR, sorry!

[zingo]

Thanks for checking 🙏