fix bug

modeling_sam_hq.py

@@ -1152,9 +1152,6 @@ class SamVisionEncoder(nn.Module):
         all_self_attentions = () if output_attentions else None
 
         for i, layer_module in enumerate(self.layers):
-            if output_hidden_states:
-                all_hidden_states = all_hidden_states + (hidden_states,)
-
             if self.gradient_checkpointing and self.training:
                 layer_outputs = self._gradient_checkpointing_func(
                     layer_module.__call__,
@@ -1164,6 +1161,8 @@ class SamVisionEncoder(nn.Module):
                 layer_outputs = layer_module(hidden_states, output_attentions=output_attentions)
 
             hidden_states = layer_outputs[0]
+            if output_hidden_states and layer_module.window_size == 0:
+                all_hidden_states = all_hidden_states + (hidden_states,)
 
             if output_attentions:
                 all_self_attentions = all_self_attentions + (layer_outputs[1],)