rm space for pure Chinese

demos/style_fs2/style_syn.py

@@ -36,7 +36,7 @@ def evaluate(args, fastspeech2_config, pwg_config):
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     with open(args.phones_dict, "r") as f:

paddlespeech/t2s/exps/fastspeech2/inference.py

@@ -84,7 +84,7 @@ def main():
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     for utt_id, sentence in sentences:

paddlespeech/t2s/exps/fastspeech2/multi_spk_synthesize_e2e.py

@@ -39,7 +39,7 @@ def evaluate(args, fastspeech2_config, pwg_config):
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     with open(args.phones_dict, "r") as f:

paddlespeech/t2s/exps/fastspeech2/synthesize_e2e.py

@@ -42,7 +42,7 @@ def evaluate(args, fastspeech2_config, pwg_config):
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     with open(args.phones_dict, "r") as f:

paddlespeech/t2s/exps/fastspeech2/synthesize_e2e_melgan.py

@@ -42,7 +42,7 @@ def evaluate(args, fastspeech2_config, melgan_config):
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     with open(args.phones_dict, "r") as f:

paddlespeech/t2s/exps/speedyspeech/inference.py

@@ -89,7 +89,7 @@ def main():
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     for utt_id, sentence in sentences:

paddlespeech/t2s/exps/speedyspeech/synthesize_e2e.py

@@ -42,7 +42,7 @@ def evaluate(args, speedyspeech_config, pwg_config):
         for line in f:
             items = line.strip().split()
             utt_id = items[0]
-            sentence = ",".join(items[1:])
+            sentence = "".join(items[1:])
             sentences.append((utt_id, sentence))
 
     with open(args.phones_dict, "r") as f:

paddlespeech/t2s/frontend/zh_frontend.py

@@ -129,6 +129,8 @@ class Frontend():
                 # we discriminate i, ii and iii
                 if c and c not in self.punc:
                     phones.append(c)
+                if c and c in self.punc:
+                    phones.append('sp')
                 if v and v not in self.punc:
                     phones.append(v)
             # add sp between sentence (replace the last punc with sp)

paddlespeech/t2s/frontend/zh_normalization/text_normlization.py

@@ -64,6 +64,8 @@ class TextNormalizer():
         List[str]
             Sentences.
         """
+        # Only for pure Chinese here
+        text = text.replace(" ", "")
         text = self.SENTENCE_SPLITOR.sub(r'\1\n', text)
         text = text.strip()
         sentences = [sentence.strip() for sentence in re.split(r'\n+', text)]