Skip to content

fix where self out contraint to make a, b numerical #11240

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
facebook-github-bot merged 1 commit into from
May 30, 2025
Merged

fix where self out contraint to make a, b numerical #11240

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
54 changes: 32 additions & 22 deletions backends/cadence/utils/facto_util.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,8 +20,8 @@
MAX_CASES = 50


def apply_tensor_contraints(op_name: str, tensor_constraints: list[object]) -> None:
additional_tensor_constraints = [
def apply_tensor_contraints(op_name: str, index: int) -> list[object]:
tensor_constraints = [
cp.Dtype.In(lambda deps: [torch.int, torch.float]),
cp.Dtype.NotIn(lambda deps: [torch.int64, torch.float64]),
cp.Value.Ge(lambda deps, dtype, struct: -(2**4)),
Expand All @@ -33,17 +33,28 @@ def apply_tensor_contraints(op_name: str, tensor_constraints: list[object]) -> N

match op_name:
case "where.self":
additional_tensor_constraints = [
cp.Dtype.In(lambda deps: [torch.float, torch.int, torch.bool]),
cp.Dtype.NotIn(lambda deps: [torch.int64, torch.float64]),
cp.Value.Ge(lambda deps, dtype, struct: -(2**4)),
cp.Value.Le(lambda deps, dtype, struct: 2**4),
cp.Rank.Ge(lambda deps: 1),
cp.Size.Ge(lambda deps, r, d: 1),
cp.Size.Le(lambda deps, r, d: 2**9),
]
if index == 0: # condition
tensor_constraints = [
cp.Dtype.In(lambda deps: [torch.bool]),
cp.Dtype.NotIn(lambda deps: [torch.int64, torch.float64]),
cp.Value.Ge(lambda deps, dtype, struct: -(2**4)),
cp.Value.Le(lambda deps, dtype, struct: 2**4),
cp.Rank.Ge(lambda deps: 1),
cp.Size.Ge(lambda deps, r, d: 1),
cp.Size.Le(lambda deps, r, d: 2**9),
]
else:
tensor_constraints = [
cp.Dtype.In(lambda deps: [torch.float, torch.int]),
cp.Dtype.NotIn(lambda deps: [torch.int64, torch.float64]),
cp.Value.Ge(lambda deps, dtype, struct: -(2**4)),
cp.Value.Le(lambda deps, dtype, struct: 2**4),
cp.Rank.Ge(lambda deps: 1),
cp.Size.Ge(lambda deps, r, d: 1),
cp.Size.Le(lambda deps, r, d: 2**9),
]
case "sigmoid.default":
additional_tensor_constraints.extend(
tensor_constraints.extend(
[
cp.Dtype.In(lambda deps: [torch.float]),
cp.Rank.Le(lambda deps: 2**2),
Expand All @@ -52,7 +63,7 @@ def apply_tensor_contraints(op_name: str, tensor_constraints: list[object]) -> N
]
)
case "rsqrt.default":
additional_tensor_constraints.extend(
tensor_constraints.extend(
[
cp.Dtype.In(lambda deps: [torch.float]),
cp.Rank.Le(lambda deps: 2**2),
Expand All @@ -63,35 +74,35 @@ def apply_tensor_contraints(op_name: str, tensor_constraints: list[object]) -> N
]
)
case "mean.dim":
additional_tensor_constraints.extend(
tensor_constraints.extend(
[
cp.Dtype.In(lambda deps: [torch.float]),
cp.Rank.Le(lambda deps: 2**2),
]
)
case "exp.default":
additional_tensor_constraints.extend(
tensor_constraints.extend(
[
cp.Rank.Le(lambda deps: 2**3),
cp.Value.Ge(lambda deps, dtype, struct: -(2**2)),
cp.Value.Le(lambda deps, dtype, struct: 2**2),
]
)
case "slice_copy.Tensor":
additional_tensor_constraints.extend(
tensor_constraints.extend(
[
cp.Rank.Le(lambda deps: 2),
cp.Value.Ge(lambda deps, dtype, struct: 1),
cp.Value.Le(lambda deps, dtype, struct: 2),
]
)
case _:
additional_tensor_constraints.extend(
tensor_constraints.extend(
[
cp.Rank.Le(lambda deps: 2**2),
]
)
tensor_constraints.extend(additional_tensor_constraints)
return tensor_constraints


def apply_scalar_contraints(op_name: str) -> list[ScalarDtype]:
Expand All @@ -107,9 +118,6 @@ def apply_scalar_contraints(op_name: str) -> list[ScalarDtype]:
def facto_testcase_gen(op_name: str) -> List[Tuple[List[str], OrderedDict[str, str]]]:
# minimal example to test add.Tensor using FACTO
spec = SpecDictDB[op_name]
tensor_constraints = []
# common tensor constraints
apply_tensor_contraints(op_name, tensor_constraints)

for index, in_spec in enumerate(copy.deepcopy(spec.inspec)):
if in_spec.type.is_scalar():
Expand Down Expand Up @@ -142,7 +150,9 @@ def facto_testcase_gen(op_name: str) -> List[Tuple[List[str], OrderedDict[str, s
]
)
elif in_spec.type.is_tensor():
spec.inspec[index].constraints.extend(tensor_constraints)
spec.inspec[index].constraints.extend(
apply_tensor_contraints(op_name, index)
)
elif in_spec.type.is_dim_list():
spec.inspec[index].constraints.extend(
[
Expand Down
Loading