Open chumaroz opened 3 years ago
It works for me
(tensorflow_venv) ➜ city-simulator git:(master) ✗ python
Python 3.8.8 (default, Feb 26 2021, 09:10:27)
[Clang 12.0.0 (clang-1200.0.32.29)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
np.isnan(np.nan)
>>> np.isnan(np.nan)
False
It works for me
(tensorflow_venv) ➜ city-simulator git:(master) ✗ python Python 3.8.8 (default, Feb 26 2021, 09:10:27) [Clang 12.0.0 (clang-1200.0.32.29)] on darwin Type "help", "copyright", "credits" or "license" for more information. >>> import numpy as np np.isnan(np.nan) >>> np.isnan(np.nan) False
@hannesa2 Yes that just proves the bug exists. And that behaviour can cause multiple issues in dependent modules downstream...
Your example works for me with conda installation https://github.com/apple/tensorflow_macos/issues/153 .
numpy version 1.20.1
This is my conda installation
I can confirm that numpy is broken for me as well
np.version.version '1.18.5'
I can confirm that the numpy that comes with the script is broken. I can reproduce the above error. It also breaks the pandas.describe() and causes seg fault.
Reproducing code example:
Error message:
the above code results in False. In other words, nan values are not recognized as nan.
NumPy/Python version information:
1.18.5 3.8.8 (default, Feb 26 2021, 09:10:27) [Clang 12.0.0 (clang-1200.0.32.29)]