thanks for the hint, virtualenv looks like an interesting option, however in my case I need to rely on several components that are already installed in the VM in Azure, including tensorflow, etc. If I use virtualenv, do I need to start from scratch?
In addition, I am not sure this will solve my problem: all I have seen is that the error code changes depending on the PYTHONPATH value. Perhaps it is a bug in the application code? 2018-05-05 9:12 GMT+02:00 dieter <die...@handshake.de>: > joseph pareti <joeparet...@gmail.com> writes: > > thank you for the advice: depending on the value of PYTHONPATH (version > > 2.7, 3.5, or unset), tensorflow stops with 3 different error traps > > "PYTHONPATH" usually is used when you have private Python modules > not installed at the standard place. Nowadays, you can use > a so called "virtual environment" for your local installations. > Look whether "virtualenv" (or "venv") is installed in your environment. > It can be used to create a virtual environment. > Then install your own modules in this virtual environment -- > ensuring not to mix things for Python 2 and Python 3. > Finally unset "PYTHONPATH" and run your Python in the virtual environment. > > -- > https://mail.python.org/mailman/listinfo/python-list > -- https://mail.python.org/mailman/listinfo/python-list
