Skip to content

Avoid log(0) in iou_loss, which will lead to Nan. #733

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 2 commits into from
Jun 11, 2019
Merged

Avoid log(0) in iou_loss, which will lead to Nan. #733

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
90f859b
Update losses.py
Joker316701882 May 29, 2019
112c513
Update losses.py
Joker316701882 May 29, 2019
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
6 changes: 3 additions & 3 deletions mmdet/core/loss/losses.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -214,7 +214,7 @@ def weighted_iou_loss(pred,
loss = bounded_iou_loss(
pred[inds], target[inds], beta=beta, eps=eps, reduction='sum')
else:
loss = iou_loss(pred[inds], target[inds], reduction='sum')
loss = iou_loss(pred[inds], target[inds], eps=eps, reduction='sum')
loss = loss[None] / avg_factor
return loss

Expand Down Expand Up @@ -248,8 +248,8 @@ def _expand_binary_labels(labels, label_weights, label_channels):
return bin_labels, bin_label_weights


def iou_loss(pred_bboxes, target_bboxes, reduction='mean'):
ious = bbox_overlaps(pred_bboxes, target_bboxes, is_aligned=True)
def iou_loss(pred_bboxes, target_bboxes, eps=1e-6, reduction='mean'):
ious = bbox_overlaps(pred_bboxes, target_bboxes, is_aligned=True) + eps
loss = -ious.log()

reduction_enum = F._Reduction.get_enum(reduction)
Expand Down