Various machine learning backends benefit greatly by running with GPU support. We can already run Keras with a TensorFlow backend on DSx but it's CPU only which makes it unusably slow for many situations. It would be great to allow DSx to take advantage of 1 or more GPUs while training a network.
Why is it useful?
|Who would benefit from this IDEA?|
How should it work?
|Submitter Tags||deep learning|