BrainSlug: Transparent Acceleration of Deep Learning Through Depth-First Parallelism

Open in new window