Adding specific versions to librosa and numba

[johannesmols]

Newer packages of those libraries have changed things and the methods used in the code are no longer available. Setting specific versions in the requirements file avoids installing those versions.

[seungwonpark]

Thanks! Could you please elaborate on what kind of librosa functions are not working with recent librosa versions?

[enm10k]

Thanks! Could you please elaborate on what kind of librosa functions are not working with recent librosa versions?

The same problem is reproduced in my environment. The error says AttributeError: module 'librosa' has no attribute 'output'.

...
'LibriSpeech/train-clean-360/2628', 'LibriSpeech/train-clean-360/3972']
/home/ec2-user/voicefilter/utils/hparams.py:18: YAMLLoadWarning: calling yaml.load_all() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
  for doc in docs:
  0%|          | 0/100000 [00:00<?, ?it/s]
multiprocessing.pool.RemoteTraceback: 
"""
Traceback (most recent call last):
  File "/usr/lib64/python3.7/multiprocessing/pool.py", line 121, in worker
    result = (True, func(*args, **kwds))
  File "generator.py", line 144, in train_wrapper
    mix(hp, args, audio, num, s1_dvec, s1_target, s2, train=True)
  File "generator.py", line 65, in mix
    librosa.output.write_wav(target_wav_path, w1, srate)
AttributeError: module 'librosa' has no attribute 'output'
"""
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "generator.py", line 154, in <module>
    r = list(tqdm.tqdm(p.imap(train_wrapper, arr), total=len(arr)))
  File "/home/ec2-user/.venv/voicefilter/lib64/python3.7/site-packages/tqdm/std.py", line 1170, in __iter__
    for obj in iterable:
  File "/usr/lib64/python3.7/multiprocessing/pool.py", line 748, in next
    raise value
AttributeError: module 'librosa' has no attribute 'output'

[ycpss91255]

This is the environment of my experiment, maybe x86_64 is a bit incomplete, but you can refer to it.

x86_64

• llvmlite==0.31.0
• tqdm==4.52.0
• numpy
• torch==1.6.0
• torchvision==0.7.0
• pyyaml
• librosa==0.6.3
• mir_eval
• matplotlib
• tensorboardX
• Pillow
• ffmpeg-normalize
• numba==0.48
• SoundFile==0.10.2

---

ppc64le

• pip==21.0.1
• setuptools==54.1.2
• wheel==0.36.2
• numpy==1.20.1
• llvmlite==0.31.0
• tqdm==4.36.1
• pybind11
• PyYAML==5.4.1
• Pillow==7.1.2
• numba==0.48.0
• scipy==1.6.1
• scikit-learn==0.21.1
• librosa==0.6.0
• mir_eval
• matplotlib
• tensorboardX
• ffmpeg-normalize
• SoundFile
• torch==1.3.1
• torchvision==0.4.0