blackteay/GPT-SoVITS
1
0
Fork 0
mirror of https://github.com/RVC-Boss/GPT-SoVITS.git synced 2025-04-06 03:57:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

chores

Browse Source
This commit is contained in:
XXXXRT666 2024-08-22 23:28:16 +08:00
parent 03a43edcd0
commit 1fca1ec17c
4 changed files with 11 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
GPT_SoVITS/TTS_infer_pack/TTS.py
View File

@ -182,6 +182,10 @@ class TTS_Config:
def _load_configs(self, configs_path: str)->dict: def _load_configs(self, configs_path: str)->dict:
if os.path.exists(configs_path):
...
else:
self.save_configs()
with open(configs_path, 'r') as f: with open(configs_path, 'r') as f:
configs = yaml.load(f, Loader=yaml.FullLoader) configs = yaml.load(f, Loader=yaml.FullLoader)
@ -748,7 +752,8 @@ class TTS:
phones, bert_features, norm_text = \ phones, bert_features, norm_text = \
self.text_preprocessor.segment_and_extract_feature_for_text( self.text_preprocessor.segment_and_extract_feature_for_text(
prompt_text, prompt_text,
prompt_lang) prompt_lang,
self.configs.version)
self.prompt_cache["phones"] = phones self.prompt_cache["phones"] = phones
self.prompt_cache["bert_features"] = bert_features self.prompt_cache["bert_features"] = bert_features
self.prompt_cache["norm_text"] = norm_text self.prompt_cache["norm_text"] = norm_text
@ -760,7 +765,7 @@ class TTS:
t1 = ttime() t1 = ttime()
data:list = None data:list = None
if not return_fragment: if not return_fragment:
data = self.text_preprocessor.preprocess(text, text_lang, text_split_method) data = self.text_preprocessor.preprocess(text, text_lang, text_split_method, self.configs.version)
if len(data) == 0: if len(data) == 0:
yield self.configs.sampling_rate, np.zeros(int(self.configs.sampling_rate), yield self.configs.sampling_rate, np.zeros(int(self.configs.sampling_rate),
dtype=np.int16) dtype=np.int16)

6
GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py
View File

@ -55,9 +55,9 @@ class TextPreprocessor:
self.tokenizer = tokenizer self.tokenizer = tokenizer
self.device = device self.device = device
def preprocess(self, text:str, lang:str, text_split_method:str, version:str="v1")->List[Dict]: def preprocess(self, text:str, lang:str, text_split_method:str, version:str="v2")->List[Dict]:
print(i18n("############ 切分文本 ############")) print(i18n("############ 切分文本 ############"))
text = self.replace_consecutive_punctuation(text) # 变量命名应该是写错了 text = self.replace_consecutive_punctuation(text)
texts = self.pre_seg_text(text, lang, text_split_method) texts = self.pre_seg_text(text, lang, text_split_method)
result = [] result = []
print(i18n("############ 提取文本Bert特征 ############")) print(i18n("############ 提取文本Bert特征 ############"))
@ -204,7 +204,7 @@ class TextPreprocessor:
phone_level_feature = torch.cat(phone_level_feature, dim=0) phone_level_feature = torch.cat(phone_level_feature, dim=0)
return phone_level_feature.T return phone_level_feature.T
def clean_text_inf(self, text:str, language:str, version:str="v1"): def clean_text_inf(self, text:str, language:str, version:str="v2"):
phones, word2ph, norm_text = clean_text(text, language, version) phones, word2ph, norm_text = clean_text(text, language, version)
phones = cleaned_text_to_sequence(phones, version) phones = cleaned_text_to_sequence(phones, version)
return phones, word2ph, norm_text return phones, word2ph, norm_text

1
GPT_SoVITS/configs/.gitignore vendored Normal file
View File

@ -0,0 +1 @@
*.yaml

24
GPT_SoVITS/configs/tts_infer.yaml
View File

@ -1,24 +0,0 @@
custom:
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
device: cuda
is_half: true
t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
version: v2
vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
default:
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
device: cpu
is_half: false
t2s_weights_path: GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
version: v1
vits_weights_path: GPT_SoVITS/pretrained_models/s2G488k.pth
default_v2:
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
device: cpu
is_half: false
t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
version: v2
vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth