3

我尝试使用 librosa 将原始音频转换为 mel-spectrogram 并收到此错误。一开始，它运行得很好，但是在我安装了tensorflow和keras之后，它就出错了。我已卸载 tensorflow 并重试，但仍然出现此错误。

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-1-7d05a4b64ba1> in <module>()
      1 import matplotlib.pyplot as plt
----> 2 import librosa
      3 import librosa.display
      4 import os
      5 import numpy as np

D:\download\WinPython-64bit-3.6.3.0Qt5\python-3.6.3.amd64\lib\site-packages\librosa\__init__.py in <module>()
     12 from . import core
     13 from . import beat
---> 14 from . import decompose
     15 from . import effects
     16 from . import feature

D:\download\WinPython-64bit-3.6.3.0Qt5\python-3.6.3.amd64\lib\site-packages\librosa\decompose.py in <module>()
     17 from scipy.ndimage import median_filter
     18 
---> 19 import sklearn.decomposition
     20 
     21 from . import core

...

D:\download\WinPython-64bit-3.6.3.0Qt5\python-3.6.3.amd64\lib\site-packages\sklearn\metrics\cluster\unsupervised.py in <module>()
     10 from ...utils import check_random_state
     11 from ...utils import check_X_y
---> 12 from ..pairwise import pairwise_distances
     13 from ...preprocessing import LabelEncoder
     14 

D:\download\WinPython-64bit-3.6.3.0Qt5\python-3.6.3.amd64\lib\site-packages\sklearn\metrics\pairwise.py in <module>()
     28 from ..externals.joblib import cpu_count
     29 
---> 30 from .pairwise_fast import _chi2_kernel_fast, _sparse_manhattan
     31 
     32 

ImportError: DLL load failed: The specified module could not be found.

我认为这是由 lib sklearn 的模块成对引起的，但我不确定。这是我的代码

import matplotlib.pyplot as plt
import librosa
import librosa.display
import os
import numpy as np

TRAIN_DATA_DIR = "genres/blues/"
num_files = 0
list_files = []
for filename in os.listdir(TRAIN_DATA_DIR):
    if (filename.endswith(".au")):
        num_files+=1
        list_files.append(filename)
        print(filename)
        fullpath = os.path.join(TRAIN_DATA_DIR, filename)
        y, sr = librosa.load(fullpath)
        librosa.feature.melspectrogram(y=y, sr=sr)
        D = np.abs(librosa.stft(y))**2
        S = librosa.feature.melspectrogram(S=D)
        S = librosa.feature.melspectrogram(y=y, sr=sr, n_mels=128, max=8000)

我已经从许多来源进行了搜索，但没有找到希望。请帮我。我怎样才能解决这个问题？谢谢

4

0 回答 0