bacow/GPT-SoVITS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

more code refactor

Browse Source
This commit is contained in:
Blaise
2024-01-16 17:14:18 +01:00
parent 0d92575115
commit 0d3d47f3c3
44 changed files with 4516 additions and 2623 deletions
Download Patch File Download Diff File Expand all files Collapse all files

104
GPT_SoVITS/text/english.py
View File

@@ -8,20 +8,87 @@ from string import punctuation
from text import symbols
current_file_path = os.path.dirname(__file__)
CMU_DICT_PATH = os.path.join(current_file_path, 'cmudict.rep')
CACHE_PATH = os.path.join(current_file_path, 'cmudict_cache.pickle')
CMU_DICT_PATH = os.path.join(current_file_path, "cmudict.rep")
CACHE_PATH = os.path.join(current_file_path, "cmudict_cache.pickle")
_g2p = G2p()
arpa = {'AH0', 'S', 'AH1', 'EY2', 'AE2', 'EH0', 'OW2', 'UH0', 'NG', 'B', 'G', 'AY0', 'M', 'AA0', 'F', 'AO0', 'ER2', 'UH1', 'IY1', 'AH2', 'DH', 'IY0', 'EY1', 'IH0', 'K', 'N', 'W', 'IY2', 'T', 'AA1', 'ER1', 'EH2', 'OY0', 'UH2', 'UW1', 'Z', 'AW2', 'AW1', 'V', 'UW2', 'AA2', 'ER', 'AW0', 'UW0', 'R', 'OW1', 'EH1', 'ZH', 'AE0', 'IH2', 'IH', 'Y', 'JH', 'P', 'AY1', 'EY0', 'OY2', 'TH', 'HH', 'D', 'ER0', 'CH', 'AO1', 'AE1', 'AO2', 'OY1', 'AY2', 'IH1', 'OW0', 'L', 'SH'}
arpa = {
"AH0",
"S",
"AH1",
"EY2",
"AE2",
"EH0",
"OW2",
"UH0",
"NG",
"B",
"G",
"AY0",
"M",
"AA0",
"F",
"AO0",
"ER2",
"UH1",
"IY1",
"AH2",
"DH",
"IY0",
"EY1",
"IH0",
"K",
"N",
"W",
"IY2",
"T",
"AA1",
"ER1",
"EH2",
"OY0",
"UH2",
"UW1",
"Z",
"AW2",
"AW1",
"V",
"UW2",
"AA2",
"ER",
"AW0",
"UW0",
"R",
"OW1",
"EH1",
"ZH",
"AE0",
"IH2",
"IH",
"Y",
"JH",
"P",
"AY1",
"EY0",
"OY2",
"TH",
"HH",
"D",
"ER0",
"CH",
"AO1",
"AE1",
"AO2",
"OY1",
"AY2",
"IH1",
"OW0",
"L",
"SH",
}
def replace_phs(phs):
rep_map = {
';': ',',
':': ',',
'\'': '-',
'"': '-'
}
rep_map = {";": ",", ":": ",", "'": "-", '"': "-"}
phs_new = []
for ph in phs:
if ph in symbols:
@@ -29,9 +96,10 @@ def replace_phs(phs):
elif ph in rep_map.keys():
phs_new.append(rep_map[ph])
else:
print('ph not in symbols: ', ph)
print("ph not in symbols: ", ph)
return phs_new
def read_dict():
g2p_dict = {}
start_line = 49
@@ -41,13 +109,13 @@ def read_dict():
while line:
if line_index >= start_line:
line = line.strip()
word_split = line.split(' ')
word_split = line.split(" ")
word = word_split[0]
syllable_split = word_split[1].split(' - ')
syllable_split = word_split[1].split(" - ")
g2p_dict[word] = []
for syllable in syllable_split:
phone_split = syllable.split(' ')
phone_split = syllable.split(" ")
g2p_dict[word].append(phone_split)
line_index = line_index + 1
@@ -57,13 +125,13 @@ def read_dict():
def cache_dict(g2p_dict, file_path):
with open(file_path, 'wb') as pickle_file:
with open(file_path, "wb") as pickle_file:
pickle.dump(g2p_dict, pickle_file)
def get_dict():
if os.path.exists(CACHE_PATH):
with open(CACHE_PATH, 'rb') as pickle_file:
with open(CACHE_PATH, "rb") as pickle_file:
g2p_dict = pickle.load(pickle_file)
else:
g2p_dict = read_dict()
@@ -71,6 +139,7 @@ def get_dict():
return g2p_dict
eng_dict = get_dict()
@@ -78,8 +147,8 @@ def text_normalize(text):
# todo: eng text normalize
return text.replace(";", ",")
def g2p(text):
def g2p(text):
phones = []
words = re.split(r"([,;.\-\?\!\s+])", text)
for w in words:
@@ -97,6 +166,7 @@ def g2p(text):
return replace_phs(phones)
if __name__ == "__main__":
# print(get_dict())
print(g2p("hello"))
@@ -106,4 +176,4 @@ if __name__ == "__main__":
# for group in syllables:
# for ph in group:
# all_phones.add(ph)
# print(all_phones)
# print(all_phones)