deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-05 18:09:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

add WeTextProcessing

Browse Source
This commit is contained in:
lyuxiang.lx
2024-07-09 23:37:54 +08:00
parent 5e97398d38
commit 7981796523
4 changed files with 12 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
cosyvoice/cli/frontend.py
View File

@@ -21,13 +21,13 @@ import torchaudio.compliance.kaldi as kaldi
import torchaudio
import os
import inflect
from tn.chinese.normalizer import Normalizer as ZhNormalizer
from tn.english.normalizer import Normalizer as EnNormalizer
try:
import ttsfrd
use_ttsfrd = True
except:
print("failed to import ttsfrd, please normalize input text manually")
except ImportError:
print("failed to import ttsfrd, use WeTextProcessing instead")
from tn.chinese.normalizer import Normalizer as ZhNormalizer
from tn.english.normalizer import Normalizer as EnNormalizer
use_ttsfrd = False
from cosyvoice.utils.frontend_utils import contains_chinese, replace_blank, replace_corner_mark, remove_bracket, spell_out_number, split_paragraph
@@ -64,8 +64,8 @@ class CosyVoiceFrontEnd:
self.frd.enable_pinyin_mix(True)
self.frd.set_breakmodel_index(1)
else:
self.zh_tn_model = ZhNormalizer(remove_erhua=False,full_to_half=False)
self.en_tn_model = EnNormalizer()
self.zh_tn_model = ZhNormalizer(remove_erhua=False, full_to_half=False)
self.en_tn_model = EnNormalizer()
def _extract_text_token(self, text):
text_token = self.tokenizer.encode(text, allowed_special=self.allowed_special)
@@ -103,7 +103,7 @@ class CosyVoiceFrontEnd:
if self.use_ttsfrd:
text = self.frd.get_frd_extra_info(text, 'input')
else:
text = self.zh_tn_model.normalize(text)
text = self.zh_tn_model.normalize(text)
text = text.replace("\n", "")
text = replace_blank(text)
text = replace_corner_mark(text)

1
cosyvoice/dataset/processor.py
View File

@@ -22,7 +22,6 @@ from torch.nn.utils.rnn import pad_sequence
import torch.nn.functional as F
torchaudio.set_audio_backend('soundfile')
torchaudio.utils.sox_utils.set_buffer_size(16500)
AUDIO_FORMAT_SETS = set(['flac', 'mp3', 'm4a', 'ogg', 'opus', 'wav', 'wma'])