Hi Venkatesh,
It is probably best to do this kind of pip install in a separate
conda environment. Doing so will reduce the chances of you breaking
your root/base environment, which is what it looks like has happened
here. Now if `conda` is broken that a big problem. I can't think
of how a pip-install command could break conda. A preferred
sequence if you start from scratch (delete everything, including
Anaconda) would be to do:
conda update conda
conda create -n tfgpu anaconda=5 python=3.6
source activate tfgpu
pip install --ignore-installed --upgrade tensorflow-gpu
That will keep all your Tensorflow pieces together and, more
importantly, AWAY from your base conda/Anaconda installation.
If you want to recover your current situation you should look at
environment variables and see what commands like `which python` and
`which conda` return.
If you would like more input from this mailing list it always helps
to include the output of commands that have failed plus some light
debugging info (environments, software versions: `conda info -a`, if
it works, provides pretty good information).
Cheers,
Ian