Update prepare.sh (#1601)

k2-fsa · Apr 20, 2024 · 9f8f0bc · 9f8f0bc
1 parent ed6bc20
commit 9f8f0bc
Showing 1 changed file with 9 additions and 0 deletions.
diff --git a/egs/multi_zh-hans/ASR/prepare.sh b/egs/multi_zh-hans/ASR/prepare.sh
@@ -299,6 +299,15 @@ if [ $stage -le 12 ] && [ $stop_stage -ge 12 ]; then
     log "Compute KeSpeech fbank for test/dev"
     ./local/compute_fbank_kespeech_dev_test.py
 
+    if [ ! -f data/fbank/kespeech/kespeech-asr_cuts_train_phase1.jsonl.gz ]; then
+      pieces=$(find data/fbank/kespeech/train_phase1_split_${num_splits} -name "kespeech-asr_cuts_train_phase1.*.jsonl.gz")
+      lhotse combine $pieces data/fbank/kespeech/kespeech-asr_cuts_train_phase1.jsonl.gz
+    fi
+    if [ ! -f data/fbank/kespeech/kespeech-asr_cuts_train_phase2.jsonl.gz ]; then
+      pieces=$(find data/fbank/kespeech/train_phase2_split_${num_splits} -name "kespeech-asr_cuts_train_phase2.*.jsonl.gz")
+      lhotse combine $pieces data/fbank/kespeech/kespeech-asr_cuts_train_phase2.jsonl.gz
+    fi
+
     touch data/fbank/.kespeech.done
   fi
 fi