Skip to content

Adjust tolerance for quantized XNN conv1d tests #8093

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

Open
GregoryComer wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Adjust tolerance for quantized XNN conv1d tests #8093

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
10 changes: 7 additions & 3 deletions backends/xnnpack/test/ops/test_conv1d.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,9 +122,13 @@ def _test_conv1d(
# For some tests we want to skip to_executorch because otherwise it will require the
# quantized operators to be loaded and we don't want to do that in the test.
if not skip_to_executorch:
tester.to_executorch().serialize().run_method_and_compare_outputs(
num_runs=10, atol=0.01, rtol=0.01
)
tester.to_executorch().serialize()
Copy link
Contributor

@digantdesai digantdesai Jan 31, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Might be ok but feels like sliding under the rug 😄

As an alternative, does it make sense to mark it as flaky (I know this is not the correct use of the flaky tag but..) so the unit-tests won't block CI and create a task to look into it?

Copy link
Contributor

@mcr229 mcr229 Jan 31, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

agreed

if quantized:
tester.run_method_and_compare_outputs(
num_runs=10, atol=0.025, rtol=0.01
)
else:
tester.run_method_and_compare_outputs()

def test_fp16_conv1d(self):
inputs = (torch.randn(2, 2, 4).to(torch.float16),)
Expand Down
Loading