运行pre.py时报错如下，数据集的.npy文件没有生成导致的，但不知道是什么原因
DennisJcy opened this issue · comments
(torch12cu113) PS E:\project\MockingBird> python .\pre.py E:\project\MockingBird
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
Using data from:
E:\project\MockingBird\aishell3\train\wav
aishell3: 0%| | 0/174 [00:00<?, ?speakers/s]Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
D:\Miniconda\envs\torch12cu113\lib\site-packages\librosa\util\decorators.py:88: UserWarning: PySoundFile failed. Trying audioread instead.
return f(*args, **kwargs)
aishell3: 0%| | 0/174 [00:05<?, ?speakers/s]
multiprocessing.pool.RemoteTraceback:
"""
Traceback (most recent call last):
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\librosa\core\audio.py", line 164, in load
y, sr_native = __soundfile_load(path, offset, duration, dtype)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\librosa\core\audio.py", line 195, in __soundfile_load
context = sf.SoundFile(path)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\soundfile.py", line 658, in init
self._file = self._open(file, mode_int, closefd)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\soundfile.py", line 1216, in _open
raise LibsndfileError(err, prefix="Error opening {0!r}: ".format(self.name))
soundfile.LibsndfileError: Error opening 'E:\project\MockingBird\SV2TTS\synthesizer\audio\audio-SSB00050001.wav_00.npy': System error.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "D:\Miniconda\envs\torch12cu113\lib\multiprocessing\pool.py", line 125, in worker
result = (True, func(*args, **kwds))
File "E:\project\MockingBird\models\synthesizer\preprocess_audio.py", line 124, in preprocess_general
wav, text = _split_on_silences(wav_fpath, words, hparams)
File "E:\project\MockingBird\models\synthesizer\preprocess_audio.py", line 88, in _split_on_silences
wav, _ = librosa.load(wav_fpath, sr= hparams.sample_rate)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\librosa\util\decorators.py", line 88, in inner_f
return f(*args, **kwargs)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\librosa\core\audio.py", line 170, in load
y, sr_native = __audioread_load(path, offset, duration, dtype)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\librosa\core\audio.py", line 226, in _audioread_load
reader = audioread.audio_open(path)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\audioread_init.py", line 127, in audio_open
return BackendClass(path)
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\audioread\rawread.py", line 59, in init
self._fh = open(filename, 'rb')
FileNotFoundError: [Errno 2] No such file or directory: 'E:\project\MockingBird\SV2TTS\synthesizer\audio\audio-SSB00050001.wav_00.npy'
"""
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "E:\project\MockingBird\pre.py", line 72, in
preprocess_dataset(**vars(args))
File "E:\project\MockingBird\models\synthesizer\preprocess.py", line 86, in preprocess_dataset
for speaker_metadata in tqdm(job, dataset, len(speaker_dirs), unit="speakers"):
File "D:\Miniconda\envs\torch12cu113\lib\site-packages\tqdm\std.py", line 1178, in iter
for obj in iterable:
File "D:\Miniconda\envs\torch12cu113\lib\multiprocessing\pool.py", line 870, in next
raise value
FileNotFoundError: [Errno 2] No such file or directory: 'E:\project\MockingBird\SV2TTS\synthesizer\audio\audio-SSB00050001.wav_00.npy'
作者您好，您最新发布的pre.py应该是有bug，我用了旧的fock仓就可以正常跑
看下linux的文档，在pre.py头部加入两行代码。
感谢回复,明白你的意思,但我是win11下运行的 @seatidespro
@DennisJcy 我也碰到了这个问题，困扰了一两天了，你解决了没有呀？
作者您好，您最新发布的pre.py应该是有bug，我用了旧的fock仓就可以正常跑
=>你是用哪个版本的仓库呀
看下linux的文档,加上两行就行了
看下linux的文档,加上两行就行了
你说的是这两行代码吧：
import torch
torch.multiprocessing.set_start_method('spawn', force=True)
我加上去以后还是报错，问题并未解决。
看下linux的文档,加上两行就行了
你说的是这两行代码吧： import torch torch.multiprocessing.set_start_method('spawn', force=True)
我加上去以后还是报错，问题并未解决。
我也报了这个错误，尽管我是WIN10专业版，也尝试在头部加了这两行代码，依然报版主上面的错误。
请问你解决了吗？@liulangdeyeshou
换旧版本吧
解决了吗，我也遇到这个问题 Linux
解决了吗，我也遇到这个问题 Linux
@huanglang0114 linux 在pre.py头部+如下两行代码不行吗？
import torch
torch.multiprocessing.set_start_method('spawn', force=True)
我是Windows 加了不行，上面有大佬建议换其他分支，还没验证
解决了吗，我也遇到这个问题 Linux
@huanglang0114 linux 在pre.py头部+如下两行代码不行吗？
import torch torch.multiprocessing.set_start_method('spawn', force=True)我是Windows 加了不行，上面有大佬建议换其他分支，还没验证
不行 不知道啥原因 我在尝试用0.0.1那个分支了
解决了吗，我也遇到这个问题 Linux
@huanglang0114 linux 在pre.py头部+如下两行代码不行吗？
import torch torch.multiprocessing.set_start_method('spawn', force=True)我是Windows 加了不行，上面有大佬建议换其他分支，还没验证
不行 不知道啥原因 我在尝试用0.0.1那个分支了
请问你解决了吗？我Win11遇到同样问题
我解决了，希望对各位有帮助，原因是：preprocess_audio.py文件中的preprocess_general函数有问题 注意观察_split_on_silences和_process_utterance的具体参数内容，注意修改
_process_utterance
请问具体要怎么改呢？
@Danny0802 , 我也遇到相同狀況，我嘗試了 #988 的 fix 之後問題解決了。