From 7dac47ca957e05bee52ea6c64fcf6d2efb403a9f Mon Sep 17 00:00:00 2001 From: XXXXRT666 <157766680+XXXXRT666@users.noreply.github.com> Date: Fri, 23 Aug 2024 17:47:27 +0800 Subject: [PATCH] chores (#1528) * chores * ... * Add files via upload * ... * remove gradio warnings * Update inference_webui.py Fix inference_cli issue --- GPT_SoVITS/TTS_infer_pack/TTS.py | 10 ++++++++-- GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py | 6 +++--- GPT_SoVITS/configs/.gitignore | 1 + GPT_SoVITS/inference_webui.py | 8 +++++++- GPT_SoVITS/inference_webui_fast.py | 5 +++++ tools/subfix_webui.py | 5 +++++ tools/uvr5/webui.py | 5 +++++ 7 files changed, 34 insertions(+), 6 deletions(-) create mode 100644 GPT_SoVITS/configs/.gitignore diff --git a/GPT_SoVITS/TTS_infer_pack/TTS.py b/GPT_SoVITS/TTS_infer_pack/TTS.py index 393c6d8..c677d77 100644 --- a/GPT_SoVITS/TTS_infer_pack/TTS.py +++ b/GPT_SoVITS/TTS_infer_pack/TTS.py @@ -182,6 +182,11 @@ class TTS_Config: def _load_configs(self, configs_path: str)->dict: + if os.path.exists(configs_path): + ... + else: + print(i18n("路径不存在,使用默认配置")) + self.save_configs(configs_path) with open(configs_path, 'r') as f: configs = yaml.load(f, Loader=yaml.FullLoader) @@ -748,7 +753,8 @@ class TTS: phones, bert_features, norm_text = \ self.text_preprocessor.segment_and_extract_feature_for_text( prompt_text, - prompt_lang) + prompt_lang, + self.configs.version) self.prompt_cache["phones"] = phones self.prompt_cache["bert_features"] = bert_features self.prompt_cache["norm_text"] = norm_text @@ -760,7 +766,7 @@ class TTS: t1 = ttime() data:list = None 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: yield self.configs.sampling_rate, np.zeros(int(self.configs.sampling_rate), dtype=np.int16) diff --git a/GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py b/GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py index 9f83f48..b90bd92 100644 --- a/GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py +++ b/GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py @@ -55,9 +55,9 @@ class TextPreprocessor: self.tokenizer = tokenizer 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 ############")) - 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) result = [] 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) 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 = cleaned_text_to_sequence(phones, version) return phones, word2ph, norm_text diff --git a/GPT_SoVITS/configs/.gitignore b/GPT_SoVITS/configs/.gitignore new file mode 100644 index 0000000..2a61605 --- /dev/null +++ b/GPT_SoVITS/configs/.gitignore @@ -0,0 +1 @@ +*.yaml \ No newline at end of file diff --git a/GPT_SoVITS/inference_webui.py b/GPT_SoVITS/inference_webui.py index 0bf41ba..5aff4ae 100644 --- a/GPT_SoVITS/inference_webui.py +++ b/GPT_SoVITS/inference_webui.py @@ -21,6 +21,11 @@ import LangSegment, os, re, sys, json import pdb import torch +try: + import gradio.analytics as analytics + analytics.version_check = lambda:None +except:... + version=os.environ.get("version","v2") pretrained_sovits_name=["GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth", "GPT_SoVITS/pretrained_models/s2G488k.pth"] pretrained_gpt_name=["GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt", "GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt"] @@ -392,7 +397,8 @@ def merge_short_text_in_array(texts, threshold): ##ref_wav_path+prompt_text+prompt_language+text(单个)+text_language+top_k+top_p+temperature # cache_tokens={}#暂未实现清理机制 cache= {} -def get_tts_wav(ref_wav_path, prompt_text, prompt_language, text, text_language, how_to_cut=i18n("不切"), top_k=20, top_p=0.6, temperature=0.6, ref_free = False,speed=1,if_freeze=False,inp_refs=123): +def get_tts_wav(ref_wav_path, prompt_text, prompt_language, text, text_language, how_to_cut=i18n("不切"), top_k=20, top_p=0.6, temperature=0.6, ref_free + =False,speed=1,if_freeze=False,inp_refs=None): global cache if ref_wav_path:pass else:gr.Warning(i18n('请上传参考音频')) diff --git a/GPT_SoVITS/inference_webui_fast.py b/GPT_SoVITS/inference_webui_fast.py index 600c393..dcc2bcf 100644 --- a/GPT_SoVITS/inference_webui_fast.py +++ b/GPT_SoVITS/inference_webui_fast.py @@ -23,6 +23,11 @@ logging.getLogger("torchaudio._extension").setLevel(logging.ERROR) import pdb import torch +try: + import gradio.analytics as analytics + analytics.version_check = lambda:None +except:... + infer_ttswebui = os.environ.get("infer_ttswebui", 9872) infer_ttswebui = int(infer_ttswebui) diff --git a/tools/subfix_webui.py b/tools/subfix_webui.py index d6624d0..9ae6c7c 100644 --- a/tools/subfix_webui.py +++ b/tools/subfix_webui.py @@ -4,6 +4,11 @@ import json import os import uuid +try: + import gradio.analytics as analytics + analytics.version_check = lambda:None +except:... + import librosa import gradio as gr import numpy as np diff --git a/tools/uvr5/webui.py b/tools/uvr5/webui.py index 1a5f2cf..dc6dd02 100644 --- a/tools/uvr5/webui.py +++ b/tools/uvr5/webui.py @@ -14,6 +14,11 @@ from mdxnet import MDXNetDereverb from vr import AudioPre, AudioPreDeEcho from bsroformer import BsRoformer_Loader +try: + import gradio.analytics as analytics + analytics.version_check = lambda:None +except:... + weight_uvr5_root = "tools/uvr5/uvr5_weights" uvr5_names = [] for name in os.listdir(weight_uvr5_root):