diff --git a/timm/models/davit.py b/timm/models/davit.py
index 9ac2b164..a2a315c6 100644
--- a/timm/models/davit.py
+++ b/timm/models/davit.py
@@ -438,10 +438,12 @@ class DaViTStage(nn.Module):
         x, size = self.patch_embed(x, size)
         x : Tensor = x
         size : Tuple[int, int] = size
+        dim1 : int = size[0]
+        dim2 : int = size[1]
         if self.grad_checkpointing and not torch.jit.is_scripting():
-            x, size = checkpoint_seq(self.blocks, x, size)
+            x, size = checkpoint_seq(self.blocks, x, (dim1, dim2))
         else:
-            x, size = self.blocks(x, size)
+            x, size = self.blocks(x, (dim1, dim2))
             
         x : Tensor = x
         size : Tuple[int, int] = size    
@@ -592,7 +594,7 @@ class DaViT(nn.Module):
     
     
     
-    def forward_features(self, x):
+    def forward_features(self, x : Tensor):
         #x, sizes = self.forward_network(x)
         size: Tuple[int, int] = (x.size(2), x.size(3))
         x, size = self.stages(x, size)