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

Refactor: Format Code with Ruff and Update Deprecated G2PW Link (#2255)

Browse Source 
* ruff check --fix

* ruff format --line-length 120 --target-version py39

* Change the link for G2PW Model

* update pytorch version and colab
This commit is contained in:
XXXXRT666
2025-04-07 09:42:47 +01:00
committed by GitHub
parent 9da7e17efe
commit 53cac93589
132 changed files with 8185 additions and 6648 deletions
Download Patch File Download Diff File Expand all files Collapse all files

88
GPT_SoVITS/text/g2pw/utils.py
View File

@@ -15,6 +15,7 @@
Credits
This code is modified from https://github.com/GitYCC/g2pW
"""
import os
import re
@@ -24,14 +25,14 @@ def wordize_and_map(text: str):
index_map_from_text_to_word = []
index_map_from_word_to_text = []
while len(text) > 0:
match_space = re.match(r'^ +', text)
match_space = re.match(r"^ +", text)
if match_space:
space_str = match_space.group(0)
index_map_from_text_to_word += [None] * len(space_str)
text = text[len(space_str):]
text = text[len(space_str) :]
continue
match_en = re.match(r'^[a-zA-Z0-9]+', text)
match_en = re.match(r"^[a-zA-Z0-9]+", text)
if match_en:
en_word = match_en.group(0)
@@ -42,7 +43,7 @@ def wordize_and_map(text: str):
index_map_from_text_to_word += [len(words)] * len(en_word)
words.append(en_word)
text = text[len(en_word):]
text = text[len(en_word) :]
else:
word_start_pos = len(index_map_from_text_to_word)
word_end_pos = word_start_pos + 1
@@ -63,15 +64,14 @@ def tokenize_and_map(tokenizer, text: str):
for word, (word_start, word_end) in zip(words, word2text):
word_tokens = tokenizer.tokenize(word)
if len(word_tokens) == 0 or word_tokens == ['[UNK]']:
if len(word_tokens) == 0 or word_tokens == ["[UNK]"]:
index_map_from_token_to_text.append((word_start, word_end))
tokens.append('[UNK]')
tokens.append("[UNK]")
else:
current_word_start = word_start
for word_token in word_tokens:
word_token_len = len(re.sub(r'^##', '', word_token))
index_map_from_token_to_text.append(
(current_word_start, current_word_start + word_token_len))
word_token_len = len(re.sub(r"^##", "", word_token))
index_map_from_token_to_text.append((current_word_start, current_word_start + word_token_len))
current_word_start = current_word_start + word_token_len
tokens.append(word_token)
@@ -85,53 +85,51 @@ def tokenize_and_map(tokenizer, text: str):
def _load_config(config_path: os.PathLike):
import importlib.util
spec = importlib.util.spec_from_file_location('__init__', config_path)
spec = importlib.util.spec_from_file_location("__init__", config_path)
config = importlib.util.module_from_spec(spec)
spec.loader.exec_module(config)
return config
default_config_dict = {
'manual_seed': 1313,
'model_source': 'bert-base-chinese',
'window_size': 32,
'num_workers': 2,
'use_mask': True,
'use_char_phoneme': False,
'use_conditional': True,
'param_conditional': {
'affect_location': 'softmax',
'bias': True,
'char-linear': True,
'pos-linear': False,
'char+pos-second': True,
'char+pos-second_lowrank': False,
'lowrank_size': 0,
'char+pos-second_fm': False,
'fm_size': 0,
'fix_mode': None,
'count_json': 'train.count.json'
"manual_seed": 1313,
"model_source": "bert-base-chinese",
"window_size": 32,
"num_workers": 2,
"use_mask": True,
"use_char_phoneme": False,
"use_conditional": True,
"param_conditional": {
"affect_location": "softmax",
"bias": True,
"char-linear": True,
"pos-linear": False,
"char+pos-second": True,
"char+pos-second_lowrank": False,
"lowrank_size": 0,
"char+pos-second_fm": False,
"fm_size": 0,
"fix_mode": None,
"count_json": "train.count.json",
},
'lr': 5e-5,
'val_interval': 200,
'num_iter': 10000,
'use_focal': False,
'param_focal': {
'alpha': 0.0,
'gamma': 0.7
"lr": 5e-5,
"val_interval": 200,
"num_iter": 10000,
"use_focal": False,
"param_focal": {"alpha": 0.0, "gamma": 0.7},
"use_pos": True,
"param_pos ": {
"weight": 0.1,
"pos_joint_training": True,
"train_pos_path": "train.pos",
"valid_pos_path": "dev.pos",
"test_pos_path": "test.pos",
},
'use_pos': True,
'param_pos ': {
'weight': 0.1,
'pos_joint_training': True,
'train_pos_path': 'train.pos',
'valid_pos_path': 'dev.pos',
'test_pos_path': 'test.pos'
}
}
def load_config(config_path: os.PathLike, use_default: bool=False):
def load_config(config_path: os.PathLike, use_default: bool = False):
config = _load_config(config_path)
if use_default:
for attr, val in default_config_dict.items():