[TTS]fix open encoding (#2865)

paddlespeech/cli/tts/infer.py

@@ -292,19 +292,19 @@ class TTSExecutor(BaseExecutor):
         with open(self.voc_config) as f:
             self.voc_config = CfgNode(yaml.safe_load(f))
 
-        with open(self.phones_dict, "r") as f:
+        with open(self.phones_dict, 'rt', encoding='utf-8') as f:
             phn_id = [line.strip().split() for line in f.readlines()]
         vocab_size = len(phn_id)
 
         tone_size = None
         if self.tones_dict:
-            with open(self.tones_dict, "r") as f:
+            with open(self.tones_dict, 'rt', encoding='utf-8') as f:
                 tone_id = [line.strip().split() for line in f.readlines()]
             tone_size = len(tone_id)
 
         spk_num = None
         if self.speaker_dict:
-            with open(self.speaker_dict, 'rt') as f:
+            with open(self.speaker_dict, 'rt', encoding='utf-8') as f:
                 spk_id = [line.strip().split() for line in f.readlines()]
             spk_num = len(spk_id)
 

paddlespeech/t2s/exps/ernie_sat/synthesize_e2e.py

@@ -437,7 +437,7 @@ if __name__ == '__main__':
 
     vocab_phones = {}
 
-    with open(args.phones_dict, 'rt') as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     for phn, id in phn_id:
         vocab_phones[phn] = int(id)

paddlespeech/t2s/exps/ernie_sat/train.py

@@ -109,7 +109,7 @@ def train_sp(args, config):
         num_workers=config.num_workers)
     print("dataloaders done!")
 
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)

paddlespeech/t2s/exps/fastspeech2/train.py

@@ -67,7 +67,7 @@ def train_sp(args, config):
     if args.speaker_dict is not None:
         print("multiple speaker fastspeech2!")
         collate_fn = fastspeech2_multi_spk_batch_fn
-        with open(args.speaker_dict, 'rt') as f:
+        with open(args.speaker_dict, 'rt', encoding='utf-8') as f:
             spk_id = [line.strip().split() for line in f.readlines()]
         spk_num = len(spk_id)
         fields += ["spk_id"]
@@ -123,7 +123,7 @@ def train_sp(args, config):
         num_workers=config.num_workers)
     print("dataloaders done!")
 
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)

paddlespeech/t2s/exps/speedyspeech/synthesize_e2e.py

@@ -39,18 +39,18 @@ def evaluate(args, speedyspeech_config, pwg_config):
 
     # construct dataset for evaluation
     sentences = []
-    with open(args.text, 'rt') as f:
+    with open(args.text, 'rt', encoding='utf-8') as f:
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
             sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)
-    with open(args.tones_dict, "r") as f:
+    with open(args.tones_dict, 'rt', encoding='utf-8') as f:
         tone_id = [line.strip().split() for line in f.readlines()]
     tone_size = len(tone_id)
     print("tone_size:", tone_size)

paddlespeech/t2s/exps/speedyspeech/train.py

@@ -70,7 +70,7 @@ def train_sp(args, config):
     if args.speaker_dict is not None:
         print("multiple speaker speedyspeech!")
         collate_fn = speedyspeech_multi_spk_batch_fn
-        with open(args.speaker_dict, 'rt') as f:
+        with open(args.speaker_dict, 'rt', encoding='utf-8') as f:
             spk_id = [line.strip().split() for line in f.readlines()]
         spk_num = len(spk_id)
         fields += ["spk_id"]
@@ -133,11 +133,11 @@ def train_sp(args, config):
         collate_fn=collate_fn,
         num_workers=config.num_workers)
     print("dataloaders done!")
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)
-    with open(args.tones_dict, "r") as f:
+    with open(args.tones_dict, 'rt', encoding='utf-8') as f:
         tone_id = [line.strip().split() for line in f.readlines()]
     tone_size = len(tone_id)
     print("tone_size:", tone_size)

paddlespeech/t2s/exps/syn_utils.py

@@ -106,7 +106,7 @@ def get_chunks(data, block_size: int, pad_size: int):
 def get_sentences(text_file: Optional[os.PathLike], lang: str='zh'):
     # construct dataset for evaluation
     sentences = []
-    with open(text_file, 'rt') as f:
+    with open(text_file, 'rt', encoding='utf-8') as f:
         for line in f:
             if line.strip() != "":
                 items = re.split(r"\s+", line.strip(), 1)
@@ -325,17 +325,17 @@ def get_am_inference(am: str='fastspeech2_csmsc',
                      tones_dict: Optional[os.PathLike]=None,
                      speaker_dict: Optional[os.PathLike]=None,
                      return_am: bool=False):
-    with open(phones_dict, "r") as f:
+    with open(phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     tone_size = None
     if tones_dict is not None:
-        with open(tones_dict, "r") as f:
+        with open(tones_dict, 'rt', encoding='utf-8') as f:
             tone_id = [line.strip().split() for line in f.readlines()]
         tone_size = len(tone_id)
     spk_num = None
     if speaker_dict is not None:
-        with open(speaker_dict, 'rt') as f:
+        with open(speaker_dict, 'rt', encoding='utf-8') as f:
             spk_id = [line.strip().split() for line in f.readlines()]
         spk_num = len(spk_id)
     odim = am_config.n_mels

paddlespeech/t2s/exps/tacotron2/train.py

@@ -119,7 +119,7 @@ def train_sp(args, config):
         num_workers=config.num_workers)
     print("dataloaders done!")
 
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)

paddlespeech/t2s/exps/transformer_tts/train.py

@@ -114,7 +114,7 @@ def train_sp(args, config):
         num_workers=config.num_workers)
     print("dataloaders done!")
 
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)

paddlespeech/t2s/exps/vits/train.py

@@ -78,7 +78,7 @@ def train_sp(args, config):
     if args.speaker_dict is not None:
         print("multiple speaker vits!")
         collate_fn = vits_multi_spk_batch_fn
-        with open(args.speaker_dict, 'rt') as f:
+        with open(args.speaker_dict, 'rt', encoding='utf-8') as f:
             spk_id = [line.strip().split() for line in f.readlines()]
         spk_num = len(spk_id)
         fields += ["spk_id"]
@@ -132,7 +132,7 @@ def train_sp(args, config):
         num_workers=config.num_workers)
     print("dataloaders done!")
 
-    with open(args.phones_dict, "r") as f:
+    with open(args.phones_dict, 'rt', encoding='utf-8') as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)

paddlespeech/t2s/frontend/phonectic.py

@@ -58,7 +58,7 @@ class English(Phonetics):
         self.punc = "：，；。？！“”‘’':,;.?!"
         self.text_normalizer = TextNormalizer()
         if phone_vocab_path:
-            with open(phone_vocab_path, 'rt') as f:
+            with open(phone_vocab_path, 'rt', encoding='utf-8') as f:
                 phn_id = [line.strip().split() for line in f.readlines()]
             for phn, id in phn_id:
                 self.vocab_phones[phn] = int(id)

paddlespeech/t2s/frontend/zh_frontend.py

@@ -144,12 +144,12 @@ class Frontend():
         self.vocab_phones = {}
         self.vocab_tones = {}
         if phone_vocab_path:
-            with open(phone_vocab_path, 'rt') as f:
+            with open(phone_vocab_path, 'rt', encoding='utf-8') as f:
                 phn_id = [line.strip().split() for line in f.readlines()]
             for phn, id in phn_id:
                 self.vocab_phones[phn] = int(id)
         if tone_vocab_path:
-            with open(tone_vocab_path, 'rt') as f:
+            with open(tone_vocab_path, 'rt', encoding='utf-8') as f:
                 tone_id = [line.strip().split() for line in f.readlines()]
             for tone, id in tone_id:
                 self.vocab_tones[tone] = int(id)