ruff check --fix

GPT_SoVITS/AR/data/dataset.py

@@ -1,19 +1,17 @@
 # modified from https://github.com/yangdongchao/SoundStorm/blob/master/soundstorm/s1/AR/data/dataset.py
 # reference: https://github.com/lifeiteng/vall-e
-import pdb
-import sys
 
 # sys.path.append("/data/docker/liujing04/gpt-vits/mq-vits-s1bert_no_bert")
-import traceback, os
+import traceback
+import os
 from typing import Dict
 from typing import List
 
 import numpy as np
 import pandas as pd
-import torch, json
+import torch
 from torch.utils.data import DataLoader
 from torch.utils.data import Dataset
-from transformers import AutoTokenizer
 
 version = os.environ.get('version',None)
 

GPT_SoVITS/AR/models/t2s_lightning_module.py

@@ -1,6 +1,7 @@
 # modified from https://github.com/yangdongchao/SoundStorm/blob/master/soundstorm/s1/AR/models/t2s_lightning_module.py
 # reference: https://github.com/lifeiteng/vall-e
-import os, sys
+import os
+import sys
 
 now_dir = os.getcwd()
 sys.path.append(now_dir)

GPT_SoVITS/AR/models/t2s_lightning_module_onnx.py

@@ -1,6 +1,7 @@
 # modified from https://github.com/yangdongchao/SoundStorm/blob/master/soundstorm/s1/AR/models/t2s_lightning_module.py
 # reference: https://github.com/lifeiteng/vall-e
-import os, sys
+import os
+import sys
 
 now_dir = os.getcwd()
 sys.path.append(now_dir)

GPT_SoVITS/AR/models/t2s_model.py

@@ -9,8 +9,6 @@ from AR.models.utils import make_pad_mask, make_pad_mask_left
 from AR.models.utils import (
     topk_sampling,
     sample,
-    logits_to_probs,
-    multinomial_sample_one_no_sync,
     dpo_loss,
     make_reject_y,
     get_batch_logps
@@ -718,7 +716,7 @@ class Text2SemanticDecoder(nn.Module):
                     idx_list[batch_index] = idx
                     y_list[batch_index] = y[i, :-1]
                 
-            if not (None in idx_list):
+            if None not in idx_list:
                 stop = True
                 
             if stop:

GPT_SoVITS/AR/models/t2s_model_onnx.py

@@ -1,7 +1,6 @@
 # modified from https://github.com/yangdongchao/SoundStorm/blob/master/soundstorm/s1/AR/models/t2s_model.py
 # reference: https://github.com/lifeiteng/vall-e
 import torch
-from tqdm import tqdm
 
 from AR.modules.embedding_onnx import SinePositionalEmbedding
 from AR.modules.embedding_onnx import TokenEmbedding

GPT_SoVITS/AR/models/utils.py

@@ -130,7 +130,7 @@ def topk_sampling(logits, top_k=10, top_p=1.0, temperature=1.0):
     return token
 
 
-from typing import Optional, Tuple
+from typing import Optional
 
 
 def multinomial_sample_one_no_sync(

GPT_SoVITS/AR/modules/activation_onnx.py

@@ -11,7 +11,6 @@ from torch.nn.init import xavier_uniform_
 from torch.nn.modules.linear import NonDynamicallyQuantizableLinear
 from torch.nn.parameter import Parameter
 
-from torch.nn import functional as F
 from AR.modules.patched_mha_with_cache_onnx import multi_head_attention_forward_patched
 
 

GPT_SoVITS/AR/modules/patched_mha_with_cache.py

@@ -5,7 +5,6 @@ from torch.nn.functional import (
     _none_or_dtype,
     _in_projection_packed,
 )
-from torch.nn import functional as F
 import torch
 # Tensor = torch.Tensor
 # from typing import Callable, List, Optional, Tuple, Union

GPT_SoVITS/AR/modules/patched_mha_with_cache_onnx.py

@@ -1,9 +1,6 @@
 from torch.nn.functional import *
 from torch.nn.functional import (
-    _mha_shape_check,
     _canonical_mask,
-    _none_or_dtype,
-    _in_projection_packed,
 )
 
 def multi_head_attention_forward_patched(

GPT_SoVITS/AR/modules/scaling.py

@@ -13,12 +13,9 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-import logging
-import math
 import random
 from typing import Optional
 from typing import Tuple
-from typing import Union
 
 import torch
 import torch.nn as nn

GPT_SoVITS/BigVGAN/bigvgan.py

@@ -451,13 +451,13 @@ class BigVGAN(
         # instantiate BigVGAN using h
         if use_cuda_kernel:
             print(
-                f"[WARNING] You have specified use_cuda_kernel=True during BigVGAN.from_pretrained(). Only inference is supported (training is not implemented)!"
+                "[WARNING] You have specified use_cuda_kernel=True during BigVGAN.from_pretrained(). Only inference is supported (training is not implemented)!"
             )
             print(
-                f"[WARNING] You need nvcc and ninja installed in your system that matches your PyTorch build is using to build the kernel. If not, the model will fail to initialize or generate incorrect waveform!"
+                "[WARNING] You need nvcc and ninja installed in your system that matches your PyTorch build is using to build the kernel. If not, the model will fail to initialize or generate incorrect waveform!"
             )
             print(
-                f"[WARNING] For detail, see the official GitHub repository: https://github.com/NVIDIA/BigVGAN?tab=readme-ov-file#using-custom-cuda-kernel-for-synthesis"
+                "[WARNING] For detail, see the official GitHub repository: https://github.com/NVIDIA/BigVGAN?tab=readme-ov-file#using-custom-cuda-kernel-for-synthesis"
             )
         model = cls(h, use_cuda_kernel=use_cuda_kernel)
 
@@ -485,7 +485,7 @@ class BigVGAN(
             model.load_state_dict(checkpoint_dict["generator"])
         except RuntimeError:
             print(
-                f"[INFO] the pretrained checkpoint does not contain weight norm. Loading the checkpoint after removing weight norm!"
+                "[INFO] the pretrained checkpoint does not contain weight norm. Loading the checkpoint after removing weight norm!"
             )
             model.remove_weight_norm()
             model.load_state_dict(checkpoint_dict["generator"])

GPT_SoVITS/BigVGAN/discriminators.py

@@ -15,7 +15,7 @@ from torchaudio.transforms import Spectrogram, Resample
 from env import AttrDict
 from utils import get_padding
 import typing
-from typing import Optional, List, Union, Dict, Tuple
+from typing import List, Tuple
 
 
 class DiscriminatorP(torch.nn.Module):
@@ -508,7 +508,7 @@ class DiscriminatorCQT(nn.Module):
         self.cqtd_normalize_volume = self.cfg.get("cqtd_normalize_volume", False)
         if self.cqtd_normalize_volume:
             print(
-                f"[INFO] cqtd_normalize_volume set to True. Will apply DC offset removal & peak volume normalization in CQTD!"
+                "[INFO] cqtd_normalize_volume set to True. Will apply DC offset removal & peak volume normalization in CQTD!"
             )
 
     def get_2d_padding(

GPT_SoVITS/BigVGAN/loss.py

@@ -6,13 +6,12 @@
 
 
 import torch
-import torch.nn.functional as F
 import torch.nn as nn
 from librosa.filters import mel as librosa_mel_fn
 from scipy import signal
 
 import typing
-from typing import Optional, List, Union, Dict, Tuple
+from typing import List, Tuple
 from collections import namedtuple
 import math
 import functools

GPT_SoVITS/BigVGAN/train.py

@@ -328,7 +328,7 @@ def train(rank, a, h):
 
                 # PESQ calculation. only evaluate PESQ if it's speech signal (nonspeech PESQ will error out)
                 if (
-                    not "nonspeech" in mode
+                    "nonspeech" not in mode
                 ):  # Skips if the name of dataset (in mode string) contains "nonspeech"
 
                     # Resample to 16000 for pesq

GPT_SoVITS/TTS_infer_pack/TTS.py

@@ -1,6 +1,8 @@
 from copy import deepcopy
 import math
-import os, sys, gc
+import os
+import sys
+import gc
 import random
 import traceback
 import time
@@ -10,7 +12,7 @@ now_dir = os.getcwd()
 sys.path.append(now_dir)
 import ffmpeg
 import os
-from typing import Generator, List, Tuple, Union
+from typing import List, Tuple, Union
 import numpy as np
 import torch
 import torch.nn.functional as F
@@ -22,14 +24,13 @@ from feature_extractor.cnhubert import CNHubert
 from module.models import SynthesizerTrn, SynthesizerTrnV3
 from peft import LoraConfig, get_peft_model
 import librosa
-from time import time as ttime
 from tools.i18n.i18n import I18nAuto, scan_language_list
 from tools.my_utils import load_audio
 from module.mel_processing import spectrogram_torch
 from TTS_infer_pack.text_segmentation_method import splits
 from TTS_infer_pack.TextPreprocessor import TextPreprocessor
 from BigVGAN.bigvgan import BigVGAN
-from module.mel_processing import spectrogram_torch,mel_spectrogram_torch
+from module.mel_processing import mel_spectrogram_torch
 from process_ckpt import get_sovits_version_from_path_fast, load_sovits_new
 language=os.environ.get("language","Auto")
 language=sys.argv[-1] if sys.argv[-1] in scan_language_list() else language
@@ -250,7 +251,7 @@ class TTS_Config:
 
         self.device = self.configs.get("device", torch.device("cpu"))
         if "cuda" in str(self.device) and not torch.cuda.is_available():
-            print(f"Warning: CUDA is not available, set device to CPU.")
+            print("Warning: CUDA is not available, set device to CPU.")
             self.device = torch.device("cpu")
 
         self.is_half = self.configs.get("is_half", False)

GPT_SoVITS/TTS_infer_pack/TextPreprocessor.py

@@ -1,5 +1,6 @@
 
-import os, sys
+import os
+import sys
 import threading
 
 from tqdm import tqdm

GPT_SoVITS/download.py

@@ -1,4 +1,5 @@
-import os, sys
+import os
+import sys
 now_dir = os.getcwd()
 sys.path.insert(0, now_dir)
 from text.g2pw import G2PWPinyin

GPT_SoVITS/export_torch_script.py

@@ -3,7 +3,6 @@
 import argparse
 from typing import Optional
 from my_utils import load_audio
-from text import cleaned_text_to_sequence
 import torch
 import torchaudio
 
@@ -813,11 +812,11 @@ import json
 def export_symbel(version='v2'):
     if version=='v1':
         symbols = text._symbol_to_id_v1
-        with open(f"onnx/symbols_v1.json", "w") as file:
+        with open("onnx/symbols_v1.json", "w") as file:
             json.dump(symbols, file, indent=4)
     else:
         symbols = text._symbol_to_id_v2
-        with open(f"onnx/symbols_v2.json", "w") as file:
+        with open("onnx/symbols_v2.json", "w") as file:
             json.dump(symbols, file, indent=4)
 
 def main():

GPT_SoVITS/export_torch_script_v3.py

@@ -6,16 +6,16 @@ from export_torch_script import (
     spectrogram_torch,
 )
 from f5_tts.model.backbones.dit import DiT
-from feature_extractor import cnhubert
 from inference_webui import get_phones_and_bert
 import librosa
 from module import commons
-from module.mel_processing import mel_spectrogram_torch, spectral_normalize_torch
+from module.mel_processing import mel_spectrogram_torch
 from module.models_onnx import CFM, SynthesizerTrnV3
 import numpy as np
 import torch._dynamo.config
 import torchaudio
-import logging, uvicorn
+import logging
+import uvicorn
 import torch
 import soundfile
 from librosa.filters import mel as librosa_mel_fn
@@ -942,7 +942,7 @@ def test_():
     
     cfm.eval()
 
-    logger.info(f"cfm ok")
+    logger.info("cfm ok")
 
     dict_s1 = torch.load("GPT_SoVITS/pretrained_models/s1v3.ckpt")
     # v2 的 gpt 也可以用
@@ -957,7 +957,7 @@ def test_():
     t2s_m = torch.jit.script(t2s_m)
     t2s_m.eval()
     # t2s_m.top_k = 15
-    logger.info(f"t2s_m ok")
+    logger.info("t2s_m ok")
 
 
     vq_model: torch.jit.ScriptModule = torch.jit.load(
@@ -967,7 +967,7 @@ def test_():
     # vq_model = vq_model.half().to(device)
     vq_model.eval()
     # vq_model = sovits.vq_model
-    logger.info(f"vq_model ok")
+    logger.info("vq_model ok")
 
     # gpt_sovits_v3_half = torch.jit.load("onnx/ad/gpt_sovits_v3_half.pt")
     # gpt_sovits_v3_half = torch.jit.optimize_for_inference(gpt_sovits_v3_half)
@@ -975,7 +975,7 @@ def test_():
     # gpt_sovits_v3_half = gpt_sovits_v3_half.cuda()
     # gpt_sovits_v3_half.eval()
     gpt_sovits_v3_half = ExportGPTSovitsHalf(sovits.hps, t2s_m, vq_model)
-    logger.info(f"gpt_sovits_v3_half ok")
+    logger.info("gpt_sovits_v3_half ok")
 
     # init_bigvgan()
     # global bigvgan_model
@@ -985,7 +985,7 @@ def test_():
     bigvgan_model = bigvgan_model.cuda()
     bigvgan_model.eval()
 
-    logger.info(f"bigvgan ok")
+    logger.info("bigvgan ok")
 
     gpt_sovits_v3 = GPTSoVITSV3(gpt_sovits_v3_half, cfm, bigvgan_model)
     gpt_sovits_v3 = torch.jit.script(gpt_sovits_v3)

GPT_SoVITS/f5_tts/model/backbones/dit.py

@@ -11,7 +11,6 @@ from __future__ import annotations
 
 import torch
 from torch import nn
-import torch.nn.functional as F
 from torch.utils.checkpoint import checkpoint
 
 from x_transformers.x_transformers import RotaryEmbedding

GPT_SoVITS/feature_extractor/cnhubert.py

@@ -1,9 +1,5 @@
-import time
 
-import librosa
 import torch
-import torch.nn.functional as F
-import soundfile as sf
 import os
 from transformers import logging as tf_logging
 tf_logging.set_verbosity_error()

GPT_SoVITS/inference_webui.py

@@ -7,7 +7,9 @@
 全部按日文识别
 '''
 import logging
-import traceback,torchaudio,warnings
+import traceback
+import torchaudio
+import warnings
 logging.getLogger("markdown_it").setLevel(logging.ERROR)
 logging.getLogger("urllib3").setLevel(logging.ERROR)
 logging.getLogger("httpcore").setLevel(logging.ERROR)
@@ -18,8 +20,10 @@ logging.getLogger("torchaudio._extension").setLevel(logging.ERROR)
 logging.getLogger("multipart.multipart").setLevel(logging.ERROR)
 warnings.simplefilter(action='ignore', category=FutureWarning)
 
-import os, re, sys, json
+import os
-import pdb
+import re
+import sys
+import json
 import torch
 from text.LangSegmenter import LangSegmenter
 
@@ -42,12 +46,12 @@ for i in range(3):
 pretrained_gpt_name,pretrained_sovits_name = _
 
 
-if os.path.exists(f"./weight.json"):
+if os.path.exists("./weight.json"):
     pass
 else:
-    with open(f"./weight.json", 'w', encoding="utf-8") as file:json.dump({'GPT':{},'SoVITS':{}},file)
+    with open("./weight.json", 'w', encoding="utf-8") as file:json.dump({'GPT':{},'SoVITS':{}},file)
 
-with open(f"./weight.json", 'r', encoding="utf-8") as file:
+with open("./weight.json", 'r', encoding="utf-8") as file:
     weight_data = file.read()
     weight_data=json.loads(weight_data)
     gpt_path = os.environ.get(
@@ -87,7 +91,6 @@ from feature_extractor import cnhubert
 cnhubert.cnhubert_base_path = cnhubert_base_path
 
 from GPT_SoVITS.module.models import SynthesizerTrn,SynthesizerTrnV3
-import numpy as np
 import random
 def set_seed(seed):
     if seed == -1:
@@ -104,9 +107,8 @@ from AR.models.t2s_lightning_module import Text2SemanticLightningModule
 from text import cleaned_text_to_sequence
 from text.cleaner import clean_text
 from time import time as ttime
-from tools.my_utils import load_audio
 from tools.i18n.i18n import I18nAuto, scan_language_list
-from peft import LoraConfig, PeftModel, get_peft_model
+from peft import LoraConfig, get_peft_model
 
 language=os.environ.get("language","Auto")
 language=sys.argv[-1] if sys.argv[-1] in scan_language_list() else language
@@ -327,9 +329,8 @@ def change_gpt_weights(gpt_path):
 
 change_gpt_weights(gpt_path)
 os.environ["HF_ENDPOINT"]          = "https://hf-mirror.com"
-import torch,soundfile
+import torch
 now_dir = os.getcwd()
-import soundfile
 
 def init_bigvgan():
     global bigvgan_model

GPT_SoVITS/inference_webui_fast.py

@@ -7,7 +7,10 @@
 全部按日文识别
 '''
 import random
-import os, re, logging, json
+import os
+import re
+import logging
+import json
 import sys
 now_dir = os.getcwd()
 sys.path.append(now_dir)
@@ -20,7 +23,6 @@ logging.getLogger("httpx").setLevel(logging.ERROR)
 logging.getLogger("asyncio").setLevel(logging.ERROR)
 logging.getLogger("charset_normalizer").setLevel(logging.ERROR)
 logging.getLogger("torchaudio._extension").setLevel(logging.ERROR)
-import pdb
 import torch
 
 try:
@@ -182,12 +184,12 @@ for i in range(3):
 pretrained_gpt_name,pretrained_sovits_name = _
 
 
-if os.path.exists(f"./weight.json"):
+if os.path.exists("./weight.json"):
     pass
 else:
-    with open(f"./weight.json", 'w', encoding="utf-8") as file:json.dump({'GPT':{},'SoVITS':{}},file)
+    with open("./weight.json", 'w', encoding="utf-8") as file:json.dump({'GPT':{},'SoVITS':{}},file)
 
-with open(f"./weight.json", 'r', encoding="utf-8") as file:
+with open("./weight.json", 'r', encoding="utf-8") as file:
     weight_data = file.read()
     weight_data=json.loads(weight_data)
     gpt_path = os.environ.get(

GPT_SoVITS/module/data_utils.py

@@ -1,22 +1,13 @@
-import time
-import logging
 import os
 import random
 import traceback
-import numpy as np
 import torch
 import torch.utils.data
 from tqdm import tqdm
 
-from module import commons
 from module.mel_processing import spectrogram_torch,spec_to_mel_torch
 from text import cleaned_text_to_sequence
-from utils import load_wav_to_torch, load_filepaths_and_text
 import torch.nn.functional as F
-from functools import lru_cache
-import requests
-from scipy.io import wavfile
-from io import BytesIO
 from tools.my_utils import load_audio
 version = os.environ.get('version',None)
 # ZeroDivisionError fixed by Tybost (https://github.com/RVC-Boss/GPT-SoVITS/issues/79)

GPT_SoVITS/module/losses.py

@@ -1,7 +1,6 @@
 import math
 
 import torch
-from torch.nn import functional as F
 
 
 def feature_loss(fmap_r, fmap_g):

GPT_SoVITS/module/mel_processing.py

@@ -1,16 +1,5 @@
-import math
-import os
-import random
 import torch
-from torch import nn
-import torch.nn.functional as F
 import torch.utils.data
-import numpy as np
-import librosa
-import librosa.util as librosa_util
-from librosa.util import normalize, pad_center, tiny
-from scipy.signal import get_window
-from scipy.io.wavfile import read
 from librosa.filters import mel as librosa_mel_fn
 
 MAX_WAV_VALUE = 32768.0

GPT_SoVITS/module/models.py

@@ -1,9 +1,6 @@
 import warnings
 warnings.filterwarnings("ignore")
-import copy
 import math
-import os
-import pdb
 
 import torch
 from torch import nn
@@ -13,7 +10,7 @@ from module import commons
 from module import modules
 from module import attentions
 from f5_tts.model import DiT
-from torch.nn import Conv1d, ConvTranspose1d, AvgPool1d, Conv2d
+from torch.nn import Conv1d, ConvTranspose1d, Conv2d
 from torch.nn.utils import weight_norm, remove_weight_norm, spectral_norm
 from module.commons import init_weights, get_padding
 from module.mrte_model import MRTE
@@ -22,7 +19,8 @@ from module.quantize import ResidualVectorQuantizer
 from text import symbols as symbols_v1
 from text import symbols2 as symbols_v2
 from torch.cuda.amp import autocast
-import contextlib,random
+import contextlib
+import random
 
 
 class StochasticDurationPredictor(nn.Module):

GPT_SoVITS/module/models_onnx.py

@@ -1,4 +1,3 @@
-import copy
 import math
 from typing import Optional
 import torch
@@ -11,14 +10,13 @@ from module import attentions_onnx as attentions
 
 from f5_tts.model import DiT
 
-from torch.nn import Conv1d, ConvTranspose1d, AvgPool1d, Conv2d
+from torch.nn import Conv1d, ConvTranspose1d, Conv2d
 from torch.nn.utils import weight_norm, remove_weight_norm, spectral_norm
 from module.commons import init_weights, get_padding
 from module.quantize import ResidualVectorQuantizer
 # from text import symbols
 from text import symbols as symbols_v1
 from text import symbols2 as symbols_v2
-from torch.cuda.amp import autocast
 
 
 class StochasticDurationPredictor(nn.Module):

GPT_SoVITS/module/quantize.py

@@ -7,7 +7,6 @@
 """Residual vector quantizer implementation."""
 
 from dataclasses import dataclass, field
-import math
 import typing as tp
 
 import torch

GPT_SoVITS/onnx_export.py

@@ -10,7 +10,6 @@ cnhubert.cnhubert_base_path = cnhubert_base_path
 ssl_model = cnhubert.get_model()
 from text import cleaned_text_to_sequence
 import soundfile
-from tools.my_utils import load_audio
 import os
 import json
 

GPT_SoVITS/prepare_datasets/1-get-text.py

@@ -14,13 +14,10 @@ bert_pretrained_dir = os.environ.get("bert_pretrained_dir")
 import torch
 is_half = eval(os.environ.get("is_half", "True")) and torch.cuda.is_available()
 version = os.environ.get('version', None)
-import sys, numpy as np, traceback, pdb
+import traceback
 import os.path
-from glob import glob
-from tqdm import tqdm
 from text.cleaner import clean_text
 from transformers import AutoModelForMaskedLM, AutoTokenizer
-import numpy as np
 from tools.my_utils import clean_path
 
 # inp_text=sys.argv[1]

GPT_SoVITS/prepare_datasets/2-get-hubert-wav32k.py

@@ -1,6 +1,7 @@
 # -*- coding: utf-8 -*-
 
-import sys,os
+import sys
+import os
 inp_text=                           os.environ.get("inp_text")
 inp_wav_dir=                        os.environ.get("inp_wav_dir")
 exp_name=                           os.environ.get("exp_name")
@@ -14,7 +15,8 @@ cnhubert.cnhubert_base_path=                os.environ.get("cnhubert_base_dir")
 import torch
 is_half = eval(os.environ.get("is_half", "True")) and torch.cuda.is_available()
 
-import pdb,traceback,numpy as np,logging
+import traceback
+import numpy as np
 from scipy.io import wavfile
 import librosa
 now_dir = os.getcwd()

GPT_SoVITS/prepare_datasets/3-get-semantic.py

@@ -26,17 +26,13 @@ else:
     version = "v3"
 import torch
 is_half = eval(os.environ.get("is_half", "True")) and torch.cuda.is_available()
-import math, traceback
+import traceback
-import multiprocessing
+import sys
-import sys, pdb
 
 now_dir = os.getcwd()
 sys.path.append(now_dir)
-from random import shuffle
+import logging
-import torch.multiprocessing as mp
+import utils
-from glob import glob
-from tqdm import tqdm
-import logging, librosa, utils
 if version!="v3":
     from module.models import SynthesizerTrn
 else:

GPT_SoVITS/process_ckpt.py

@@ -1,7 +1,8 @@
 import traceback
 from collections import OrderedDict
 from time import time as ttime
-import shutil,os
+import shutil
+import os
 import torch
 from tools.i18n.i18n import I18nAuto
 

GPT_SoVITS/s1_train.py

@@ -1,6 +1,5 @@
 # modified from https://github.com/feng-yufei/shared_debugging_code/blob/main/train_t2s.py
 import os
-import pdb
 
 if "_CUDA_VISIBLE_DEVICES" in os.environ:
     os.environ["CUDA_VISIBLE_DEVICES"] = os.environ["_CUDA_VISIBLE_DEVICES"]
@@ -8,7 +7,8 @@ import argparse
 import logging
 from pathlib import Path
 
-import torch, platform
+import torch
+import platform
 from pytorch_lightning import seed_everything
 from pytorch_lightning import Trainer
 from pytorch_lightning.callbacks import ModelCheckpoint
@@ -24,8 +24,6 @@ torch.set_float32_matmul_precision("high")
 from AR.utils import get_newest_ckpt
 
 from collections import OrderedDict
-from time import time as ttime
-import shutil
 from process_ckpt import my_save
 
 

GPT_SoVITS/s2_train.py

@@ -1,6 +1,7 @@
 import warnings
 warnings.filterwarnings("ignore")
-import utils, os
+import utils
+import os
 hps = utils.get_hparams(stage=2)
 os.environ["CUDA_VISIBLE_DEVICES"] = hps.train.gpu_numbers.replace("-", ",")
 import torch
@@ -8,11 +9,11 @@ from torch.nn import functional as F
 from torch.utils.data import DataLoader
 from torch.utils.tensorboard import SummaryWriter
 import torch.multiprocessing as mp
-import torch.distributed as dist, traceback
+import torch.distributed as dist
 from torch.nn.parallel import DistributedDataParallel as DDP
 from torch.cuda.amp import autocast, GradScaler
 from tqdm import tqdm
-import logging, traceback
+import logging
 
 logging.getLogger("matplotlib").setLevel(logging.INFO)
 logging.getLogger("h5py").setLevel(logging.INFO)

GPT_SoVITS/s2_train_v3.py

@@ -1,18 +1,18 @@
 import warnings
 warnings.filterwarnings("ignore")
-import utils, os
+import utils
+import os
 hps = utils.get_hparams(stage=2)
 os.environ["CUDA_VISIBLE_DEVICES"] = hps.train.gpu_numbers.replace("-", ",")
 import torch
-from torch.nn import functional as F
 from torch.utils.data import DataLoader
 from torch.utils.tensorboard import SummaryWriter
 import torch.multiprocessing as mp
-import torch.distributed as dist, traceback
+import torch.distributed as dist
 from torch.nn.parallel import DistributedDataParallel as DDP
 from torch.cuda.amp import autocast, GradScaler
 from tqdm import tqdm
-import logging, traceback
+import logging
 
 logging.getLogger("matplotlib").setLevel(logging.INFO)
 logging.getLogger("h5py").setLevel(logging.INFO)
@@ -27,10 +27,7 @@ from module.data_utils import (
 )
 from module.models import (
     SynthesizerTrnV3 as SynthesizerTrn,
-    MultiPeriodDiscriminator,
 )
-from module.losses import generator_loss, discriminator_loss, feature_loss, kl_loss
-from module.mel_processing import mel_spectrogram_torch, spec_to_mel_torch
 from process_ckpt import savee
 
 torch.backends.cudnn.benchmark = False

GPT_SoVITS/s2_train_v3_lora.py

@@ -1,25 +1,25 @@
 import warnings
 warnings.filterwarnings("ignore")
-import utils, os
+import utils
+import os
 hps = utils.get_hparams(stage=2)
 os.environ["CUDA_VISIBLE_DEVICES"] = hps.train.gpu_numbers.replace("-", ",")
 import torch
-from torch.nn import functional as F
 from torch.utils.data import DataLoader
 from torch.utils.tensorboard import SummaryWriter
 import torch.multiprocessing as mp
-import torch.distributed as dist, traceback
+import torch.distributed as dist
 from torch.nn.parallel import DistributedDataParallel as DDP
 from torch.cuda.amp import autocast, GradScaler
 from tqdm import tqdm
-import logging, traceback
+import logging
 
 logging.getLogger("matplotlib").setLevel(logging.INFO)
 logging.getLogger("h5py").setLevel(logging.INFO)
 logging.getLogger("numba").setLevel(logging.INFO)
 from random import randint
 from module import commons
-from peft import LoraConfig, PeftModel, get_peft_model
+from peft import LoraConfig, get_peft_model
 from module.data_utils import (
     TextAudioSpeakerLoaderV3 as TextAudioSpeakerLoader,
     TextAudioSpeakerCollateV3 as TextAudioSpeakerCollate,
@@ -27,10 +27,7 @@ from module.data_utils import (
 )
 from module.models import (
     SynthesizerTrnV3 as SynthesizerTrn,
-    MultiPeriodDiscriminator,
 )
-from module.losses import generator_loss, discriminator_loss, feature_loss, kl_loss
-from module.mel_processing import mel_spectrogram_torch, spec_to_mel_torch
 from process_ckpt import savee
 from collections import OrderedDict as od
 torch.backends.cudnn.benchmark = False

GPT_SoVITS/text/cantonese.py

@@ -1,6 +1,5 @@
 # reference: https://huggingface.co/spaces/Naozumi0512/Bert-VITS2-Cantonese-Yue/blob/main/text/chinese.py
 
-import sys
 import re
 import cn2an
 import ToJyutping

GPT_SoVITS/text/chinese.py

@@ -1,5 +1,4 @@
 import os
-import pdb
 import re
 
 import cn2an
@@ -17,7 +16,8 @@ pinyin_to_symbol_map = {
     for line in open(os.path.join(current_file_path, "opencpop-strict.txt")).readlines()
 }
 
-import jieba_fast, logging
+import jieba_fast
+import logging
 jieba_fast.setLogLevel(logging.CRITICAL)
 import jieba_fast.posseg as psg
 

GPT_SoVITS/text/chinese2.py

@@ -1,10 +1,9 @@
 import os
-import pdb
 import re
 
 import cn2an
 from pypinyin import lazy_pinyin, Style
-from pypinyin.contrib.tone_convert import to_normal, to_finals_tone3, to_initials, to_finals
+from pypinyin.contrib.tone_convert import to_finals_tone3, to_initials
 
 from text.symbols import punctuation
 from text.tone_sandhi import ToneSandhi
@@ -18,7 +17,8 @@ pinyin_to_symbol_map = {
     for line in open(os.path.join(current_file_path, "opencpop-strict.txt")).readlines()
 }
 
-import jieba_fast, logging
+import jieba_fast
+import logging
 jieba_fast.setLogLevel(logging.CRITICAL)
 import jieba_fast.posseg as psg
 

GPT_SoVITS/text/english.py

@@ -8,7 +8,6 @@ from text.symbols import punctuation
 
 from text.symbols2 import symbols
 
-import unicodedata
 from builtins import str as unicode
 from text.en_normalization.expend import normalize
 from nltk.tokenize import TweetTokenizer

GPT_SoVITS/text/g2pw/onnx_api.py

@@ -5,7 +5,8 @@ import warnings
 warnings.filterwarnings("ignore")
 import json
 import os
-import zipfile,requests
+import zipfile
+import requests
 from typing import Any
 from typing import Dict
 from typing import List

GPT_SoVITS/text/japanese.py

@@ -58,7 +58,7 @@ try:
 
     if os.path.exists(USERDIC_BIN_PATH):
         pyopenjtalk.update_global_jtalk_with_user_dict(USERDIC_BIN_PATH)   
-except Exception as e:
+except Exception:
     # print(e)
     import pyopenjtalk
     # failed to load user dictionary, ignore.

GPT_SoVITS/text/korean.py

@@ -16,7 +16,7 @@ if os.name == 'nt':
             spam_spec = importlib.util.find_spec("eunjeon")
             non_found = spam_spec is None
             if non_found:
-                print(f'you have to install eunjeon. install it...')
+                print('you have to install eunjeon. install it...')
             else:
                 installpath = spam_spec.submodule_search_locations[0]
                 if not (re.match(r'^[A-Za-z0-9_/\\:.\-]*$', installpath)):

GPT_SoVITS/text/symbols.py

@@ -1,4 +1,3 @@
-import os
 
 # punctuation = ['!', '?', '…', ",", ".","@"]#@是SP停顿
 punctuation = ["!", "?", "…", ",", "."]  # @是SP停顿

GPT_SoVITS/text/symbols2.py

@@ -1,4 +1,3 @@
-import os
 
 # punctuation = ['!', '?', '…', ",", ".","@"]#@是SP停顿
 punctuation = ["!", "?", "…", ",", "."]  # @是SP停顿

GPT_SoVITS/utils.py

@@ -9,9 +9,7 @@ import traceback
 
 import librosa
 import numpy as np
-from scipy.io.wavfile import read
 import torch
-import logging
 
 logging.getLogger("numba").setLevel(logging.ERROR)
 logging.getLogger("matplotlib").setLevel(logging.ERROR)
@@ -132,7 +130,6 @@ def plot_spectrogram_to_numpy(spectrogram):
         mpl_logger = logging.getLogger("matplotlib")
         mpl_logger.setLevel(logging.WARNING)
     import matplotlib.pylab as plt
-    import numpy as np
 
     fig, ax = plt.subplots(figsize=(10, 2))
     im = ax.imshow(spectrogram, aspect="auto", origin="lower", interpolation="none")
@@ -158,7 +155,6 @@ def plot_alignment_to_numpy(alignment, info=None):
         mpl_logger = logging.getLogger("matplotlib")
         mpl_logger.setLevel(logging.WARNING)
     import matplotlib.pylab as plt
-    import numpy as np
 
     fig, ax = plt.subplots(figsize=(6, 4))
     im = ax.imshow(

api.py

@@ -142,7 +142,8 @@ RESP: 无
 
 
 import argparse
-import os,re
+import os
+import re
 import sys
 
 now_dir = os.getcwd()
@@ -152,10 +153,11 @@ sys.path.append("%s/GPT_SoVITS" % (now_dir))
 import signal
 from text.LangSegmenter import LangSegmenter
 from time import time as ttime
-import torch, torchaudio
+import torch
+import torchaudio
 import librosa
 import soundfile as sf
-from fastapi import FastAPI, Request, Query, HTTPException
+from fastapi import FastAPI, Request, Query
 from fastapi.responses import StreamingResponse, JSONResponse
 import uvicorn
 from transformers import AutoModelForMaskedLM, AutoTokenizer
@@ -163,12 +165,11 @@ import numpy as np
 from feature_extractor import cnhubert
 from io import BytesIO
 from module.models import SynthesizerTrn, SynthesizerTrnV3
-from peft import LoraConfig, PeftModel, get_peft_model
+from peft import LoraConfig, get_peft_model
 from AR.models.t2s_lightning_module import Text2SemanticLightningModule
 from text import cleaned_text_to_sequence
 from text.cleaner import clean_text
 from module.mel_processing import spectrogram_torch
-from tools.my_utils import load_audio
 import config as global_config
 import logging
 import subprocess
@@ -221,7 +222,7 @@ def resample(audio_tensor, sr0):
     return resample_transform_dict[sr0](audio_tensor)
 
 
-from module.mel_processing import spectrogram_torch,mel_spectrogram_torch
+from module.mel_processing import mel_spectrogram_torch
 spec_min = -12
 spec_max = 2
 def norm_spec(x):
@@ -860,7 +861,7 @@ def handle(refer_wav_path, prompt_text, prompt_language, text, text_language, cu
         if not default_refer.is_ready():
             return JSONResponse({"code": 400, "message": "未指定参考音频且接口无预设"}, status_code=400)
 
-    if not sample_steps in [4,8,16,32]:
+    if sample_steps not in [4,8,16,32]:
         sample_steps = 32
 
     if cut_punc == None:
@@ -990,10 +991,10 @@ logger.info(f"编码格式: {media_type}")
 # 音频数据类型
 if args.sub_type.lower() == 'int32':
     is_int32 = True
-    logger.info(f"数据类型: int32")
+    logger.info("数据类型: int32")
 else:
     is_int32 = False
-    logger.info(f"数据类型: int16")
+    logger.info("数据类型: int16")
 
 # 初始化模型
 cnhubert.cnhubert_base_path = cnhubert_base_path

api_v2.py

@@ -112,15 +112,13 @@ import wave
 import signal
 import numpy as np
 import soundfile as sf
-from fastapi import FastAPI, Request, HTTPException, Response
+from fastapi import FastAPI, Response
 from fastapi.responses import StreamingResponse, JSONResponse
-from fastapi import FastAPI, UploadFile, File
 import uvicorn
 from io import BytesIO
 from tools.i18n.i18n import I18nAuto
 from GPT_SoVITS.TTS_infer_pack.TTS import TTS, TTS_Config
 from GPT_SoVITS.TTS_infer_pack.text_segmentation_method import get_method_names as get_cut_method_names
-from fastapi.responses import StreamingResponse
 from pydantic import BaseModel
 # print(sys.path)
 i18n = I18nAuto()
@@ -337,7 +335,7 @@ async def tts_handle(req:dict):
             audio_data = pack_audio(BytesIO(), audio_data, sr, media_type).getvalue()
             return Response(audio_data, media_type=f"audio/{media_type}")
     except Exception as e:
-        return JSONResponse(status_code=400, content={"message": f"tts failed", "Exception": str(e)})
+        return JSONResponse(status_code=400, content={"message": "tts failed", "Exception": str(e)})
     
 
 
@@ -415,7 +413,7 @@ async def set_refer_aduio(refer_audio_path: str = None):
     try:
         tts_pipeline.set_ref_audio(refer_audio_path)
     except Exception as e:
-        return JSONResponse(status_code=400, content={"message": f"set refer audio failed", "Exception": str(e)})
+        return JSONResponse(status_code=400, content={"message": "set refer audio failed", "Exception": str(e)})
     return JSONResponse(status_code=200, content={"message": "success"})
 
 
@@ -444,7 +442,7 @@ async def set_gpt_weights(weights_path: str = None):
             return JSONResponse(status_code=400, content={"message": "gpt weight path is required"})
         tts_pipeline.init_t2s_weights(weights_path)
     except Exception as e:
-        return JSONResponse(status_code=400, content={"message": f"change gpt weight failed", "Exception": str(e)})
+        return JSONResponse(status_code=400, content={"message": "change gpt weight failed", "Exception": str(e)})
 
     return JSONResponse(status_code=200, content={"message": "success"})
 
@@ -456,7 +454,7 @@ async def set_sovits_weights(weights_path: str = None):
             return JSONResponse(status_code=400, content={"message": "sovits weight path is required"})
         tts_pipeline.init_vits_weights(weights_path)
     except Exception as e:
-        return JSONResponse(status_code=400, content={"message": f"change sovits weight failed", "Exception": str(e)})
+        return JSONResponse(status_code=400, content={"message": "change sovits weight failed", "Exception": str(e)})
     return JSONResponse(status_code=200, content={"message": "success"})
 
 
@@ -466,7 +464,7 @@ if __name__ == "__main__":
         if host == 'None':   # 在调用时使用 -a None 参数，可以让api监听双栈
             host = None
         uvicorn.run(app=APP, host=host, port=port, workers=1)
-    except Exception as e:
+    except Exception:
         traceback.print_exc()
         os.kill(os.getpid(), signal.SIGTERM)
         exit(0)

config.py

@@ -1,4 +1,5 @@
-import sys,os
+import sys
+import os
 
 import torch
 

gpt-sovits_kaggle.ipynb

@@ -143,11 +143,7 @@
     "# 开启推理页面\n",
     "%cd /kaggle/working/GPT-SoVITS/\n",
     "!npm install -g localtunnel\n",
-    "import subprocess\n",
     "import threading\n",
-    "import time\n",
-    "import socket\n",
-    "import urllib.request\n",
     "def iframe_thread(port):\n",
     "    while True:\n",
     "        time.sleep(0.5)\n",

tools/AP_BWE_main/models/model.py

@@ -1,7 +1,7 @@
 import torch
 import torch.nn.functional as F
 import torch.nn as nn
-from torch.nn.utils import weight_norm, remove_weight_norm, spectral_norm
+from torch.nn.utils import weight_norm, spectral_norm
 # from utils import init_weights, get_padding
 def get_padding(kernel_size, dilation=1):
     return int((kernel_size*dilation - dilation)/2)

tools/audio_sr.py

@@ -1,24 +1,15 @@
 from __future__ import absolute_import, division, print_function, unicode_literals
-import sys,os
+import sys
-import traceback
+import os
 AP_BWE_main_dir_path=os.path.join(os.path.dirname(os.path.abspath(__file__)), 'AP_BWE_main')
 sys.path.append(AP_BWE_main_dir_path)
-import glob
-import argparse
 import json
-from re import S
 import torch
-import numpy as np
-import torchaudio
-import time
 import torchaudio.functional as aF
 # from attrdict import AttrDict####will be bug in py3.10
 
 from datasets1.dataset import amp_pha_stft, amp_pha_istft
 from models.model import APNet_BWE_Model
-import soundfile as sf
-import matplotlib.pyplot as plt
-from rich.progress import track
 
 class AP_BWE():
     def __init__(self,device,DictToAttrRecursive,checkpoint_file=None):

tools/cmd-denoise.py

@@ -1,4 +1,5 @@
-import os,argparse
+import os
+import argparse
 import traceback
 
 from modelscope.pipelines import pipeline

tools/i18n/scan_i18n.py

@@ -116,7 +116,7 @@ def update_i18n_json(json_file, standard_keys):
     if num_miss_translation > 0:
         print(f"\033[31m{'[Failed] Missing Translation'.ljust(KEY_LEN)}: {num_miss_translation}\033[0m")
     else:
-        print(f"\033[32m[Passed] All Keys Translated\033[0m")
+        print("\033[32m[Passed] All Keys Translated\033[0m")
     # 将处理后的结果写入 JSON 文件
     with open(json_file, "w", encoding="utf-8") as f:
         json.dump(json_data, f, ensure_ascii=False, indent=4, sort_keys=SORT_KEYS)

tools/my_utils.py

@@ -1,4 +1,5 @@
-import platform,os,traceback
+import os
+import traceback
 import ffmpeg
 import numpy as np
 import gradio as gr
@@ -21,7 +22,7 @@ def load_audio(file, sr):
             .output("-", format="f32le", acodec="pcm_f32le", ac=1, ar=sr)
             .run(cmd=["ffmpeg", "-nostdin"], capture_stdout=True, capture_stderr=True)
         )
-    except Exception as e:
+    except Exception:
         traceback.print_exc()
         raise RuntimeError(i18n("音频加载失败"))
 

tools/slice_audio.py

@@ -1,4 +1,6 @@
-import os,sys,numpy as np
+import os
+import sys
+import numpy as np
 import traceback
 from scipy.io import wavfile
 # parent_directory = os.path.dirname(os.path.abspath(__file__))

tools/slicer2.py

@@ -249,7 +249,7 @@ def main():
         soundfile.write(
             os.path.join(
                 out,
-                f"%s_%d.wav"
+                "%s_%d.wav"
                 % (os.path.basename(args.audio).rsplit(".", maxsplit=1)[0], i),
             ),
             chunk,

tools/subfix_webui.py

@@ -1,7 +1,7 @@
-import argparse,os
+import argparse
+import os
 import copy
 import json
-import os
 import uuid
 
 try:
@@ -70,7 +70,7 @@ def b_change_index(index, batch):
             # )
             {
                 "__type__": "update",
-                "label": f"Text",
+                "label": "Text",
                 "value": ""
             }
         )

tools/uvr5/bs_roformer/attend.py

@@ -54,7 +54,7 @@ class Attend(nn.Module):
 
         # similarity
 
-        sim = einsum(f"b h i d, b h j d -> b h i j", q, k) * scale
+        sim = einsum("b h i d, b h j d -> b h i j", q, k) * scale
 
         # attention
 
@@ -63,6 +63,6 @@ class Attend(nn.Module):
 
         # aggregate values
 
-        out = einsum(f"b h i j, b h j d -> b h i d", attn, v)
+        out = einsum("b h i j, b h j d -> b h i d", attn, v)
 
         return out

tools/uvr5/bs_roformer/bs_roformer.py

@@ -1,14 +1,14 @@
 from functools import partial
 
 import torch
-from torch import nn, einsum, Tensor
+from torch import nn
 from torch.nn import Module, ModuleList
 import torch.nn.functional as F
 
 from bs_roformer.attend import Attend
 from torch.utils.checkpoint import checkpoint
 
-from typing import Tuple, Optional, List, Callable
+from typing import Tuple, Optional, Callable
 # from beartype.typing import Tuple, Optional, List, Callable
 # from beartype import beartype
 

tools/uvr5/bs_roformer/mel_band_roformer.py

@@ -1,14 +1,14 @@
 from functools import partial
 
 import torch
-from torch import nn, einsum, Tensor
+from torch import nn
 from torch.nn import Module, ModuleList
 import torch.nn.functional as F
 
 from bs_roformer.attend import Attend
 from torch.utils.checkpoint import checkpoint
 
-from typing import Tuple, Optional, List, Callable
+from typing import Tuple, Optional, Callable
 # from beartype.typing import Tuple, Optional, List, Callable
 # from beartype import beartype
 

tools/uvr5/vr.py

@@ -1,6 +1,6 @@
-import os,sys
+import os
 parent_directory = os.path.dirname(os.path.abspath(__file__))
-import logging,pdb
+import logging
 logger = logging.getLogger(__name__)
 
 import librosa

tools/uvr5/webui.py

@@ -1,13 +1,13 @@
 import os
-import traceback,gradio as gr
+import traceback
+import gradio as gr
 import logging
 from tools.i18n.i18n import I18nAuto
 from tools.my_utils import clean_path
 i18n = I18nAuto()
 
 logger = logging.getLogger(__name__)
-import librosa,ffmpeg
+import ffmpeg
-import soundfile as sf
 import torch
 import sys
 from mdxnet import MDXNetDereverb

webui.py

@@ -1,4 +1,5 @@
-import os,sys
+import os
+import sys
 if len(sys.argv)==1:sys.argv.append('v2')
 version="v1"if sys.argv[1]=="v1" else"v2"
 os.environ["version"]=version
@@ -6,7 +7,11 @@ now_dir = os.getcwd()
 sys.path.insert(0, now_dir)
 import warnings
 warnings.filterwarnings("ignore")
-import json,yaml,torch,pdb,re,shutil
+import json
+import yaml
+import torch
+import re
+import shutil
 import platform
 import psutil
 import signal
@@ -45,21 +50,18 @@ for site_packages_root in site_packages_roots:
                     % (now_dir, now_dir, now_dir, now_dir, now_dir, now_dir)
                 )
             break
-        except PermissionError as e:
+        except PermissionError:
             traceback.print_exc()
 from tools import my_utils
 import shutil
-import pdb
 import subprocess
 from subprocess import Popen
-import signal
 from config import python_exec,infer_device,is_half,exp_root,webui_port_main,webui_port_infer_tts,webui_port_uvr5,webui_port_subfix,is_share
 from tools.i18n.i18n import I18nAuto, scan_language_list
 language=sys.argv[-1] if sys.argv[-1] in scan_language_list() else "Auto"
 os.environ["language"]=language
 i18n = I18nAuto(language=language)
-from scipy.io import wavfile
+from tools.my_utils import check_for_existance, check_details
-from tools.my_utils import load_audio, check_for_existance, check_details
 from multiprocessing import cpu_count
 # os.environ['PYTORCH_ENABLE_MPS_FALLBACK'] = '1' # 当遇到mps不支持的步骤时使用cpu
 try: