I tried to install TensorFlow v2.2.0 with the update to Python 3.8.3, but I got hooked. In conclusion, if I backdate to Python 3.7.6, I can install it without any problem.
Python beginners are building an environment for studying TensorFlow packages.
OS:Windows10 Enterprise 1809 64bit Python: (Before change) 3.8.3, (After change) 3.7.6
I tried to install the tensorflow package with the following command, but I got an Error statement.
Command entered
pip install tensorflow (c1)
Output result
ERROR: Could not find a version that satisfies the requirement TensorFlow (from versions: none)
ERROR: No matching distribution found for TensorFlow
I also updated pip, but the problem was not solved. Then, I downloaded the whl file from the following page and tried to pip install locally, but it still doesn't work.
https://pypi.org/project/tensorflow/#files
why? I tried various things, but I just got hooked on it.
So, after investigating various things, it is said that tensorflow may not support the Python version that has just been released. In fact, when Python v3.7 was up to date, I found a web page with a note saying "tensorflow doesn't work without Python 3.6!". Eh, really? At the above URL, there is also a file "tensorflow-2.2.0-cp38-cp38-win_amd64.whl" ...
So, quietly create a new virtual environment for Python v3.7.6 and enter the command (c1) again.
The installation started smoothly. Chan Chan.
Depending on the information, there may be a comment that "tensorflow is compatible with v3.8", so maybe v3.8.1 or v3.8.2 will support it. However, if you attack too much here, the same problem may occur with other packages, so we will proceed with v3.7.6 first.
Recommended Posts