diff --git a/ldm/modules/encoders/modules.py b/ldm/modules/encoders/modules.py
index 66f0e57..84b793f 100644
--- a/ldm/modules/encoders/modules.py
+++ b/ldm/modules/encoders/modules.py
@@ -158,6 +158,7 @@ class FrozenCLIPEmbedder(AbstractEncoder):
 
         if self.penultimate:
             z = outputs.hidden_states[-2] # simple enough
+            z = self.transformer.text_model.final_layer_norm(z)
         else:
             z = outputs.last_hidden_state