From 6982a6c1e1574ec9646581fad3cfe00f429495ac Mon Sep 17 00:00:00 2001
From: Baek-Donghyeon <71208448+nistring@users.noreply.github.com>
Date: Wed, 14 May 2025 07:34:51 +0900
Subject: [PATCH] Update dice.py

The dice loss varies with batch sizes, and this issue remains unresolved.
https://github.com/qubvel-org/segmentation_models.pytorch/issues/712#issue-1550779281
---
 segmentation_models_pytorch/losses/dice.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/segmentation_models_pytorch/losses/dice.py b/segmentation_models_pytorch/losses/dice.py
index b8baae98..38a074f2 100644
--- a/segmentation_models_pytorch/losses/dice.py
+++ b/segmentation_models_pytorch/losses/dice.py
@@ -70,7 +70,7 @@ def forward(self, y_pred: torch.Tensor, y_true: torch.Tensor) -> torch.Tensor:
 
         bs = y_true.size(0)
         num_classes = y_pred.size(1)
-        dims = (0, 2)
+        dims = (2)
 
         if self.mode == BINARY_MODE:
             y_true = y_true.view(bs, 1, -1)