After setting up machines (and helping others do it) with the suite of commonly used deep learning toolkits [1] for the N-th time, I decided it might be useful to put together a simple guide. While conceptually simple, figuring out package dependencies, getting the drivers working, et al turn out to be fairly non-trivial in practice. Until someone builds a deep learning ready Docker/VM, hope this helps a few of you save some time. Check it out: https://github.com/saiprashanths/dl-setup
This was tested on a 64 bit machine with Nvidia Titan X, running Ubuntu 14.04
Follow the discussions on Hacker News, Reddit Machine Learning
[1] Tensorflow, Caffe, Theano, Keras, Torch