Skip to content

fix: use per-sample dice computation in DiceLoss to match CrossEntropyLoss mean reduction (#74) #101

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
rtmalikian wants to merge 1 commit into
base: master
Choose a base branch
from
Open

fix: use per-sample dice computation in DiceLoss to match CrossEntropyLoss mean reduction (#74) #101

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
16 changes: 10 additions & 6 deletions code/utils/losses.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -169,12 +169,16 @@ def _one_hot_encoder(self, input_tensor):
def _dice_loss(self, score, target):
target = target.float()
smooth = 1e-5
intersect = torch.sum(score * target)
y_sum = torch.sum(target * target)
z_sum = torch.sum(score * score)
loss = (2 * intersect + smooth) / (z_sum + y_sum + smooth)
loss = 1 - loss
return loss
# Compute per-sample dice to match CrossEntropyLoss mean reduction
batch_size = score.shape[0]
loss = 0.0
for i in range(batch_size):
intersect = torch.sum(score[i] * target[i])
y_sum = torch.sum(target[i] * target[i])
z_sum = torch.sum(score[i] * score[i])
sample_loss = 1 - (2 * intersect + smooth) / (z_sum + y_sum + smooth)
loss += sample_loss
return loss / batch_size

def forward(self, inputs, target, weight=None, softmax=False):
if softmax:
Expand Down