diff --git a/models.py b/models.py
index 9e199fb..8441dea 100644
--- a/models.py
+++ b/models.py
@@ -808,6 +808,8 @@ class WindowAttention(nn.Module):
             self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1)  # Wh*Ww,Wh*Ww,nH
         relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
         relative_position_bias = 16 * torch.sigmoid(relative_position_bias)
+        print(attn.shape)
+        print(relative_position_bias.shape)
         attn = attn + relative_position_bias.unsqueeze(0)
 
         if mask is not None: