from local

egs/tedlium2/ASR/conformer_ctc3/train.py

@@ -582,6 +582,32 @@ def compute_loss(
                 use_double_scores=params.use_double_scores,
             )
 
+        elif params.condition:
+            dense_fsa_vec1 = k2.DenseFsaVec(
+                nnet_output[0],
+                supervision_segments,
+                allow_truncate=params.subsampling_factor - 1,
+            )
+            
+            dense_fsa_vec2 = k2.DenseFsaVec(
+                nnet_output[1][8],
+                supervision_segments,
+                allow_truncate=params.subsampling_factor - 1,
+            )
+
+            ctc_loss = (1-params.interctc_weight)* k2.ctc_loss(
+                decoding_graph=decoding_graph,
+                dense_fsa_vec=dense_fsa_vec1,
+                output_beam=params.beam_size,
+                reduction=params.reduction,
+                use_double_scores=params.use_double_scores,
+            ) + params.interctc_weight * k2.ctc_loss(
+                decoding_graph=decoding_graph,
+                dense_fsa_vec=dense_fsa_vec2,
+                output_beam=params.beam_size,
+                reduction=params.reduction,
+                use_double_scores=params.use_double_scores,
+            )
 
         else:
             dense_fsa_vec = k2.DenseFsaVec(