Skip to content

Widen float32 tolerance for TestRotatedBoxIou::test_iou on arm64 Linux #9636

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
Ahmed8881 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Widen float32 tolerance for TestRotatedBoxIou::test_iou on arm64 Linux #9636

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
7 changes: 7 additions & 0 deletions test/test_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
import math
import os
import platform
import sys
import time
from abc import ABC, abstractmethod
Expand Down Expand Up @@ -1889,6 +1890,12 @@ def test_iou(self, device, dtype, fmt):
pytest.xfail(
"xyxyxyxy format with float32 on CUDA has expected failure due to floating-point precision errors in format conversion"
)
elif platform.machine() in ("arm64", "aarch64"):
# Same root cause as the macOS case above: the NEON/Sleef sin/cos
# implementation used on arm64 Linux rounds slightly differently
# than the AVX backend used on x86_64, which is enough to flip
# the Graham scan's ordering in this format round-trip.
atol = 0.5
else:
atol = 1e-4
else:
Expand Down