A while ago my research lab acquired a new workstation, but my PI, well meaning as he is, purchased a system with an AMD GPU (FirePro W7100) rather than an Nvidia card, so CUDA is not an option. For the longest time I thought deep learning was not going to happen with TensorFlow using an OpenCV library, but I recently stumbled on a library PlaidML, a tensor compiler that allows for the use of OpenCL devices, and sits as a layer underneath common machine learning frameworks.
Having recently needed to convert the format of an enormous quantity of images I stumbled upon GNU parallel, a powerful and simple utility for replacing shell loops with a parallel alternative.
Mogrify with GNU Parallel Here is a simple example where I use GNU parallel to parallelize the conversion of a series of svs (a format commonly used for pathology images from digital whole slide scanners) to tiffs, a much more common format.
This is a simple list of tips and tricks for the shell that I discovered over the course of many years of messing around in shell and wanted to coalesce into a single document, partly to remember, and partly to share with others.
Persistent SSH Sessions In order to ensure that your SSH sessions do not get disconnected after a period of inactivity simply add the following to your .