Skip to content

call _upcast to consider overflow for computing giou loss #5685

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
merged 4 commits into from
Mar 28, 2022
Merged

call _upcast to consider overflow for computing giou loss #5685

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
2f8089f
call _upcast to consider overflow
developer0hye Mar 26, 2022
dc2d754
cast datatype to float
developer0hye Mar 27, 2022
e6a7769
add ":" to if
developer0hye Mar 27, 2022
ad704ed
lint Test for membership should be 'not in' (E713)
developer0hye Mar 27, 2022
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: 10 additions & 0 deletions torchvision/ops/giou_loss.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,12 @@
import torch
from torch import Tensor


def _upcast(t: Tensor) -> Tensor:
# Protects from numerical overflows in multiplications by upcasting to the equivalent higher type
if t.dtype not in (torch.float32, torch.float64):
return t.float()
return t


def generalized_box_iou_loss(
Expand Down Expand Up @@ -34,6 +42,8 @@ def generalized_box_iou_loss(
https://arxiv.org/abs/1902.09630
"""

boxes1 = _upcast(boxes1)
boxes2 = _upcast(boxes2)
x1, y1, x2, y2 = boxes1.unbind(dim=-1)
x1g, y1g, x2g, y2g = boxes2.unbind(dim=-1)

Expand Down