Skip to content

Added better error messages for type validators (stack trace) #7999

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
swolchok merged 4 commits into from
Jan 28, 2025
Merged

Added better error messages for type validators (stack trace) #7999

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
140987e
Added better error messages for type validators (stack trace)
cptspacemanspiff Jan 26, 2025
f84ed9f
changed error message format (moved newlines for better readability.)
cptspacemanspiff Jan 28, 2025
18d6312
Fixed linting issue.
cptspacemanspiff Jan 28, 2025
777c710
Fixed unit test (it checks the first element of the new pair)
cptspacemanspiff Jan 28, 2025
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
2 changes: 1 addition & 1 deletion exir/tests/test_arg_validator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,7 +64,7 @@ def forward(self, x):
ops.edge.aten._log_softmax.default.name(),
)
self.assertDictEqual(
validator.violating_ops[key],
validator.violating_ops[key][0],
{
"self": torch.bfloat16,
"__ret_0": torch.bfloat16,
Expand Down
8 changes: 4 additions & 4 deletions exir/verification/arg_validator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,9 +37,9 @@ class EdgeOpArgValidator(torch.fx.Interpreter):

def __init__(self, graph_module: torch.fx.GraphModule) -> None:
super().__init__(graph_module)
self.violating_ops: Dict[EdgeOpOverload, Dict[str, Optional[torch.dtype]]] = (
defaultdict(dict)
)
self.violating_ops: Dict[
EdgeOpOverload, Tuple[Dict[str, Optional[torch.dtype]], torch.fx.Node]
] = defaultdict(dict)

def run_node(self, n: torch.fx.Node) -> None:
self.node = n
Expand Down Expand Up @@ -125,5 +125,5 @@ def call_function( # noqa: C901 # pyre-fixme[14]

valid = target._schema.dtype_constraint.validate(tensor_arg_types)
if not valid:
self.violating_ops[target] = tensor_arg_types
self.violating_ops[target] = (tensor_arg_types, self.node)
return super().call_function(target, args, kwargs) # pyre-fixme[6]
10 changes: 8 additions & 2 deletions exir/verification/verifier.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -189,9 +189,15 @@ def _check_tensor_args_matching_op_allowed_dtype(gm: GraphModule) -> None:
return

if validator.violating_ops:
error_msg = ""
for op, node in validator.violating_ops.items():
# error_msg += f"#####################################################\n"
error_msg += f"\nOperator: {op} with args: {node[0]}\n"
error_msg += f"stack trace: {node[1].stack_trace}\n"
# error_msg += f"#####################################################\n"
raise SpecViolationError(
f"These operators are taking Tensor inputs with mismatched dtypes: {validator.violating_ops}"
"Please make sure the dtypes of the Tensor inputs are the same as the dtypes of the corresponding "
f"These operators are taking Tensor inputs with mismatched dtypes:\n{error_msg}"
"Please make sure the dtypes of the Tensor inputs are the same as the dtypes of the corresponding outputs."
)


Expand Down