init

main_s2s.sh

@@ -8,16 +8,6 @@ export LINE_NO_START=0
 export LINE_NO_END=10
 python fetch_dataset_s2s.py
 # main
-for i in $(seq 49 50);
-do
-  export N_POOL=15
-  export DATASET_ID=${i}
-  export DIRECTION="enA-koA"
-  export LINE_NO_START=$(((DATASET_ID-1) * 2500))
-  export LINE_NO_END=$((DATASET_ID * 2500))
-  echo ${LINE_NO_START}
-  python fetch_dataset_s2s.py
-done
 for i in $(seq 87 100);
 do
   export N_POOL=15
@@ -29,7 +19,7 @@ do
   python fetch_dataset_s2s.py
 done
 # tokenize
-for i in $(seq 21 36);
+for i in $(seq 37 50);
 do
   export N_POOL=15
   export DATASET_ID=${i}
@@ -339,12 +329,17 @@ export SE_MODEL="metavoice"
 export SE_MODEL="w2vbert-600m"
 export SE_MODEL="hubert-xl"
 export SE_MODEL="xlsr-2b"
-for i in $(seq 1 91);
+for i in $(seq 17 91);
 do
   export DATASET_ID=${i}
   export DIRECTION="enA-hiA"
   python attach_speaker_embedding_s2s.py
 done
+
+export SE_MODEL="hubert-xl"
+export DATASET_ID=84
+export DIRECTION="enA-hiA"
+python attach_speaker_embedding_s2s.py
 ####################
 # enA-viA: 740_598 #
 ####################
@@ -374,9 +369,9 @@ do
 done
 # speaker embedding
 export SE_MODEL="metavoice"
+export SE_MODEL="xlsr-2b"
 export SE_MODEL="w2vbert-600m"
 export SE_MODEL="hubert-xl"
-export SE_MODEL="xlsr-2b"
 for i in $(seq 1 149);
 do
   export DATASET_ID=${i}
@@ -384,7 +379,17 @@ do
   python attach_speaker_embedding_s2s.py
 done
 
+export SE_MODEL="w2vbert-600m"
+export DATASET_ID=116
+python attach_speaker_embedding_s2s.py
 
+export SE_MODEL="hubert-xl"
+for i in 107 115 126;
+do
+  export DATASET_ID=${i}
+  export DIRECTION="enA-viA"
+  python attach_speaker_embedding_s2s.py
+done
 
 ######################
 # enA-zhA: 1_289_192 #
@@ -419,10 +424,10 @@ do
 done
 # speaker embedding
 export SE_MODEL="metavoice"
-export SE_MODEL="hubert-xl"
-export SE_MODEL="xlsr-2b"
+#export SE_MODEL="xlsr-2b"
 export SE_MODEL="w2vbert-600m"
-for i in $(seq 14 258);
+export SE_MODEL="hubert-xl"
+for i in $(seq 1 258);
 do
   export DATASET_ID=${i}
   export DIRECTION="enA-zhA"
@@ -460,12 +465,28 @@ done
 # speaker embedding
 export SE_MODEL="metavoice"
 export SE_MODEL="w2vbert-600m"
-export SE_MODEL="xlsr-2b"
 export SE_MODEL="hubert-xl"
+export SE_MODEL="xlsr-2b"
 for i in $(seq 1 144);
 do
   export DATASET_ID=${i}
   export DIRECTION="enA-jaA"
   python attach_speaker_embedding_s2s.py
-  rm -rf ~/.cache/huggingface/datasets
 done
+
+export SE_MODEL="w2vbert-600m"
+for i in 9 107 126 142;
+do
+  export DATASET_ID=${i}
+  export DIRECTION="enA-jaA"
+  python attach_speaker_embedding_s2s.py
+done
+
+export SE_MODEL="hubert-xl"
+for i in 103 114 115 121 129;
+do
+  export DATASET_ID=${i}
+  export DIRECTION="enA-jaA"
+  python attach_speaker_embedding_s2s.py
+done
+

tokenize_dataset_s2s.py

@@ -26,6 +26,8 @@ def error_file(example):
             wav = audio_loader.decode_example(example[f"{side}.audio"])
             if len(wav["array"]) == 0 or len(wav["array"]) > max_seq_length:
                 return False
+        except ValueError:
+            return False
         except LibsndfileError:
             return False
     return True