I Chose to learn about tensorflow because there is a large amount of tutorials available. I’m currently working through the Udacity course which so far I am enjoying.
I choose Tensorflow, because I believe it is still the most comprehensive open source library available, although lately I have been keeping an eye on PyTorch too...
I have written image processing application from scratch (in c++) for a couple of commercial life sciences instruments. At the time, there were no open source libraries available (at least not that I was aware of).
I chose Caffe solely because I saw it referenced from the Texas Insruments AI docs I was researching for the BBAI. To date, I'm not familiar with any of them, but I figure if TI has them part of their overall AI Learning workflow, it's probably a good place to start for me.
Top Comments