Skip to content

aot_arm_compiler: Add argument to opt out of strict exports #12750

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
AdrianLundell merged 1 commit into from
Jul 23, 2025
Merged

aot_arm_compiler: Add argument to opt out of strict exports #12750

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
11 changes: 9 additions & 2 deletions examples/arm/aot_arm_compiler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -580,6 +580,13 @@ def get_args():
default="Arm/vela.ini",
help="Specify custom vela configuration file (vela.ini)",
)
parser.add_argument(
"--non_strict_export",
dest="strict_export",
required=False,
action="store_false",
help="Disable strict checking while exporting models.",
)
args = parser.parse_args()

if args.evaluate and (
Expand Down Expand Up @@ -696,7 +703,7 @@ def quantize_model(args, model: torch.nn.Module, example_inputs, compile_spec):
)
# Wrap quantized model back into an exported_program
exported_program = torch.export.export_for_training(
model_int8, example_inputs, strict=True
model_int8, example_inputs, strict=args.strict_export
)

return model_int8, exported_program
Expand Down Expand Up @@ -791,7 +798,7 @@ def transform_for_cortex_m_backend(edge):
# export_for_training under the assumption we quantize, the exported form also works
# in to_edge if we don't quantize
exported_program = torch.export.export_for_training(
model, example_inputs, strict=True
model, example_inputs, strict=args.strict_export
)
model = exported_program.module()
model_fp32 = model
Expand Down
Loading