Hi guys, I have done many experiments with MXnet . I can see that its performance beats tensorflow and pytorch with great margin but i have not experimented swift for tensorflow. Is it faster than swift for tensorflow in terms of computation?
Thanks @SibtainRazaJamali, glad you find MXNet performant! I have not done the comparison myself with the swift bindings of TF, but I don’t expect the front-end language to impact much the speed of processing of the backend if the framework is used correctly: avoid I/O bottlenecks, synchronous calls etc.