deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-05 09:59:23 +08:00
Code Issues Packages Projects Releases Wiki Activity

compatible when ttsfrd is not avaliable

Browse Source
This commit is contained in:
lyuxiang.lx
2024-07-07 13:13:42 +08:00
parent 39f02fa1f9
commit 50c7b06ea9
2 changed files with 21 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
README.md
View File

@@ -55,7 +55,10 @@ git clone https://www.modelscope.cn/iic/CosyVoice-300M-Instruct.git pretrained_m
git clone https://www.modelscope.cn/iic/CosyVoice-ttsfrd.git pretrained_models/CosyVoice-ttsfrd git clone https://www.modelscope.cn/iic/CosyVoice-ttsfrd.git pretrained_models/CosyVoice-ttsfrd
``` ```
Unzip `ttsfrd` resouce and install `ttsfrd` package Optionaly, you can unzip `ttsfrd` resouce and install `ttsfrd` package.
Notice that this step is not necessary. If you do not install `ttsfrd` package, you need to normalize input text manually.
``` sh ``` sh
cd pretrained_models/CosyVoice-ttsfrd/ cd pretrained_models/CosyVoice-ttsfrd/
unzip resource.zip -d . unzip resource.zip -d .

25
cosyvoice/cli/frontend.py
View File

@@ -21,7 +21,12 @@ import torchaudio.compliance.kaldi as kaldi
import torchaudio import torchaudio
import os import os
import inflect import inflect
import ttsfrd try:
import ttsfrd
use_ttsfrd = True
except:
print("failed to import ttsfrd, please normalize input text manually")
use_ttsfrd = False
from cosyvoice.utils.frontend_utils import contains_chinese, replace_blank, replace_corner_mark, remove_bracket, spell_out_number, split_paragraph from cosyvoice.utils.frontend_utils import contains_chinese, replace_blank, replace_corner_mark, remove_bracket, spell_out_number, split_paragraph
@@ -48,12 +53,14 @@ class CosyVoiceFrontEnd:
self.instruct = instruct self.instruct = instruct
self.allowed_special = allowed_special self.allowed_special = allowed_special
self.inflect_parser = inflect.engine() self.inflect_parser = inflect.engine()
self.frd = ttsfrd.TtsFrontendEngine() self.use_ttsfrd = use_ttsfrd
ROOT_DIR = os.path.dirname(os.path.abspath(__file__)) if self.use_ttsfrd:
assert self.frd.initialize('{}/../../pretrained_models/CosyVoice-ttsfrd/resource'.format(ROOT_DIR)) is True, 'failed to initialize ttsfrd resource' self.frd = ttsfrd.TtsFrontendEngine()
self.frd.set_lang_type('pinyin') ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
self.frd.enable_pinyin_mix(True) assert self.frd.initialize('{}/../../pretrained_models/CosyVoice-ttsfrd/resource'.format(ROOT_DIR)) is True, 'failed to initialize ttsfrd resource'
self.frd.set_breakmodel_index(1) self.frd.set_lang_type('pinyin')
self.frd.enable_pinyin_mix(True)
self.frd.set_breakmodel_index(1)
def _extract_text_token(self, text): def _extract_text_token(self, text):
text_token = self.tokenizer.encode(text, allowed_special=self.allowed_special) text_token = self.tokenizer.encode(text, allowed_special=self.allowed_special)
@@ -88,7 +95,9 @@ class CosyVoiceFrontEnd:
def text_normalize(self, text, split=True): def text_normalize(self, text, split=True):
text = text.strip() text = text.strip()
if contains_chinese(text): if contains_chinese(text):
text = self.frd.get_frd_extra_info(text, 'input').replace("\n", "") if self.use_ttsfrd:
text = self.frd.get_frd_extra_info(text, 'input')
text = text.replace("\n", "")
text = replace_blank(text) text = replace_blank(text)
text = replace_corner_mark(text) text = replace_corner_mark(text)
text = text.replace(".", "、") text = text.replace(".", "、")