I am just getting into Googles Cloud services, and have not been doing much ML in Python. I have done one or two ML projects in school using Tensorflow, but not really accustomed with the other available libraries. Is it so that if I am interested in using the Google Cloud services I should definitely be using Tensorflow, or can they be integrated with Theano or other ML libraries too? If they can be used, can they only be used in part and some functionality is lost, or are they fully available?
Just looking for some guidance if there is any argument for using something else than Tensorflow for someone in my shoes, and don't really find good guidance using Google as they promote their own product.
