I am setting up my computer to run DL with a GPU and I couldn't find info on whether one should install keras or keras-gpu. Currently I have it running with conda and keras using tensorflow-gpu as backend. What would be the difference if I switch keras to keras-gpu?
1 Answers
this is a paragraph borrowed from Wikipedia:
Keras was conceived to be an interface rather than a standalone machine-learning framework. It offers a higher-level, more intuitive set of abstractions that make it easy to develop deep learning models regardless of the computational backend used.
https://en.wikipedia.org/wiki/Keras
So as long as all three back ends of Keras (TensorFlow, Microsoft Cognitive Toolkit, or Theano) do not support Gpu versions, using keras-gpu may cause you trouble.
Related