Continuing the discussion of the last two episodes, there is one more aspect of deep learning that I would love to consider and therefore left as a full episode, that is parallelising and distributing deep learning on relatively large clusters.