diff --git a/egs/tedlium2/ASR/conformer_ctc3/.train.py.swp b/egs/tedlium2/ASR/conformer_ctc3/.train.py.swp
index 075c51b53..02369516a 100644
Binary files a/egs/tedlium2/ASR/conformer_ctc3/.train.py.swp and b/egs/tedlium2/ASR/conformer_ctc3/.train.py.swp differ
diff --git a/egs/tedlium2/ASR/conformer_ctc3/.transformer.py.swp b/egs/tedlium2/ASR/conformer_ctc3/.transformer.py.swp
index 74a458f1d..50f0cb9c7 100644
Binary files a/egs/tedlium2/ASR/conformer_ctc3/.transformer.py.swp and b/egs/tedlium2/ASR/conformer_ctc3/.transformer.py.swp differ
diff --git a/egs/tedlium2/ASR/conformer_ctc3/transformer.py b/egs/tedlium2/ASR/conformer_ctc3/transformer.py
index 4ff9062fe..6743734a6 100644
--- a/egs/tedlium2/ASR/conformer_ctc3/transformer.py
+++ b/egs/tedlium2/ASR/conformer_ctc3/transformer.py
@@ -186,14 +186,13 @@ class Transformer(nn.Module):
         encoder_memory, memory_key_padding_mask = self.run_encoder(
             x, supervision, warmup
         )
-        x = self.ctc_output(encoder_memory)
-        
         if type(encoder_memory) == tuple:
             (encoder_memory, layer_outputs) = encoder_memory
             layer_outputs = [self.ctc_output(x) for x in layer_outputs]
 
             return (x, layer_outputs), encoder_memory, memory_key_padding_mask
         else:
+            x = self.ctc_output(encoder_memory)
             return x, encoder_memory, memory_key_padding_mask
 
     def run_encoder(