Skip to content

Arm backend: Add support for masked_fill_.Scalar #16272

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Sebastian-Larsson merged 1 commit into from
Dec 16, 2025
Merged

Arm backend: Add support for masked_fill_.Scalar #16272

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions backends/arm/_passes/decompose_masked_fill_pass.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@


edge_ops = (exir_ops.edge.aten.masked_fill.Scalar,)
aten_ops = (torch.ops.aten.masked_fill.Scalar,)
aten_ops = (torch.ops.aten.masked_fill.Scalar, torch.ops.aten.masked_fill_.Scalar)


def _get_decomposition(op) -> tuple:
Expand All @@ -26,7 +26,7 @@ def _get_decomposition(op) -> tuple:
exir_ops.edge.aten.where.self,
exir_ops.edge.aten.full_like.default,
)
if op in aten_ops:
elif op in aten_ops:
return (
torch.ops.aten.where.self,
torch.ops.aten.full_like.default,
Expand All @@ -44,7 +44,7 @@ class DecomposeMaskedFillPass(ArmPass):
_passes_required_after: Set[Type[ExportPass]] = {ConvertFullLikeToFullPass}

def call_operator(self, op, args, kwargs, meta, updated=False):
if op not in (edge_ops + aten_ops):
if op not in (*aten_ops, *edge_ops):
return super().call_operator(op, args, kwargs, meta, updated)

x, mask, scalar = args
Expand Down
2 changes: 0 additions & 2 deletions backends/arm/test/models/stable_diffusion/test_CLIPTextModelWithProjection.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,9 +45,7 @@ class TestCLIPTextModelWithProjection:
"executorch_exir_dialects_edge__ops_aten_index_select_default": 1,
"executorch_exir_dialects_edge__ops_aten_slice_copy_Tensor": 1,
"executorch_exir_dialects_edge__ops_aten_view_copy_default": 1,
"executorch_exir_dialects_edge__ops_aten_where_self": 1,
"executorch_exir_dialects_edge__ops_dim_order_ops__to_dim_order_copy_default": 2,
"torch.ops.aten.scalar_tensor.default": 1,
"torch.ops.higher_order.executorch_call_delegate": 2,
}

Expand Down
46 changes: 10 additions & 36 deletions backends/arm/test/models/test_conformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,15 +72,8 @@ def test_conformer_tosa_INT():
aten_op=[],
exir_op=[],
use_to_edge_transform_and_lower=True,
)
pipeline.pop_stage("check_count.exir")
pipeline.change_args(
"run_method_and_compare_outputs",
get_test_inputs(
TestConformer.dim, TestConformer.lengths, TestConformer.num_examples
),
rtol=TestConformer.rtol,
atol=TestConformer.atol,
rtol=TestConformer.rtol,
)
pipeline.run()

Expand All @@ -93,38 +86,26 @@ def test_conformer_u55_INT():
pipeline = EthosU55PipelineINT[input_t](
TestConformer.conformer,
TestConformer.model_example_inputs,
aten_ops=TestConformer.aten_ops,
aten_ops=[],
exir_ops=[],
use_to_edge_transform_and_lower=True,
atol=TestConformer.atol,
rtol=TestConformer.rtol,
)
pipeline.change_args(
"run_method_and_compare_outputs",
get_test_inputs(
TestConformer.dim, TestConformer.lengths, TestConformer.num_examples
),
rtol=1.0,
atol=5.0,
)
pipeline.pop_stage("check_count.exir")
pipeline.run()


@common.XfailIfNoCorstone320
@pytest.mark.xfail(reason="All IO needs to have the same data type (MLETORCH-635)")
def test_conformer_u85_INT():
pipeline = EthosU85PipelineINT[input_t](
TestConformer.conformer,
TestConformer.model_example_inputs,
aten_ops=TestConformer.aten_ops,
aten_ops=[],
exir_ops=[],
use_to_edge_transform_and_lower=True,
)
pipeline.change_args(
"run_method_and_compare_outputs",
get_test_inputs(
TestConformer.dim, TestConformer.lengths, TestConformer.num_examples
),
rtol=1.0,
atol=5.0,
atol=TestConformer.atol,
rtol=TestConformer.rtol,
)
pipeline.run()

Expand All @@ -137,16 +118,9 @@ def test_conformer_vgf_quant():
aten_op=[],
exir_op=[],
use_to_edge_transform_and_lower=True,
quantize=True,
)
pipeline.pop_stage("check_count.exir")
pipeline.change_args(
"run_method_and_compare_outputs",
get_test_inputs(
TestConformer.dim, TestConformer.lengths, TestConformer.num_examples
),
rtol=TestConformer.rtol,
atol=TestConformer.atol,
rtol=TestConformer.rtol,
quantize=True,
)
pipeline.run()

Expand Down
Loading