I have followed the instructions for enabling GPU
support https://www.tensorflow.org/install/gpu?hl=ur - which installs CUDA
dependencies for GPU
to operate.
I have a native Python 3.8.5
with Tensorflow 2.4.1
, the latest release, and Python 3.7.9
with Tensorflow 2.0.0
installed in the virtual environment.
These instructions seem to be generic for all versions, because they simply install the CUDA toolkit
and some additional libraries.
However, only the native python sees the GPU
, and the one in the virtual environment does not. What needs to be done in order to enable GPU
support for the virtual environment as well? Or it is due to the fact, that this version of CUDA Toolkit
is too modern for Tensorflow 2.0.0
?
question from:
https://stackoverflow.com/questions/65883809/enabling-gpu-support-for-tensorflow-in-virtual-enviroment-for-older-version-of-t 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…