Anybody who attempts to do deep learning faces multiple choices in terms of framework. There are static graph libraries like tensorflow, theano and dynamic ones like chainer, Pytorch etc. The talk will go into details which type of library to choose based on use cases. The talk is not aimed to provide answer like what is best. The aim of the talk is to educate the audience about the tools.
There are multiple choices one needs to make while choosing a framework for doing Deep Learning. Given plethora of choices one is bound to get confused. A lot terminology to a beginner doesn't make sense. The talk will not provide you an answer. But it will detail the considerations one need to make before making a choice. We will cover all major libraries including Theano, Caffe, Caffe2, Pytorch, Tensorflow, Chainer, MxNet. Will list out the basic principles on which these libraries are based. What libraries are good for what stuff. The talk will not be about one library vs another but about educating a new practitioner so he can make an informed choice.