diff --git a/nodes/model_optimization_nodes.py b/nodes/model_optimization_nodes.py
index 4821f15..7e7e769 100644
--- a/nodes/model_optimization_nodes.py
+++ b/nodes/model_optimization_nodes.py
@@ -40,6 +40,7 @@ class BaseLoaderKJ:
                 encoder_hidden_states_mask: torch.FloatTensor = None,
                 attention_mask: Optional[torch.FloatTensor] = None,
                 image_rotary_emb: Optional[torch.Tensor] = None,
+                transformer_options={},
             ) -> Tuple[torch.Tensor, torch.Tensor]:
                 seq_txt = encoder_hidden_states.shape[1]
 
@@ -67,7 +68,7 @@ class BaseLoaderKJ:
                 joint_key = joint_key.flatten(start_dim=2)
                 joint_value = joint_value.flatten(start_dim=2)
 
-                joint_hidden_states = attention_sage(joint_query, joint_key, joint_value, self.heads, attention_mask)
+                joint_hidden_states = attention_sage(joint_query, joint_key, joint_value, self.heads, attention_mask, transformer_options=transformer_options)
 
                 txt_attn_output = joint_hidden_states[:, :seq_txt, :]
                 img_attn_output = joint_hidden_states[:, seq_txt:, :]