Swift is an interesting choice, one we haven't explored in depth. Out of curiosity, have you done any work with Swift for Tensorflow/what has your experience been?
Every time Swift for Tensorflow gets brought up I am reminded of the disappointment that they didn’t make arguably the smarter choice and do it in Julia
I have no idea what TensorFlow is (other than the basics) but I enjoyed reading that entire document because it did such a wonderful job of explaining a complex and potentially contentious decision. It’s fascinating to see Swift feature so strongly in a pragmatic analysis that doesn’t explicitly favour Apple platform interop.
Objectively it's a failure, since no one programs deep learning applications in Swift for TensorFlow, but it's cool to see that parts of it still live on.
Swift is a great all purpose language. Before Google dropped the Swift TensorFlow project, I had started writing a “Swift AI” book. I bailed on the book, and instead have a live-book GitHub repo with code and README files for the various non-TensorFlow projects (which I tossed out to /dev/null). The live book is at https://github.com/mark-watson/Swift-AI-live-book-and-code
I'd be surprised if it's dead. At least two of the people on the Swift for Tensorflow team were building a deep learning library in Swift before Swift for Tensorflow was a thing. Much of their work is a part of Swift for Tensorflow.
I remember when the creator of Swift joined google and there was a lot of noise about Swift and Tensorflow. i told people while it seemed like a nice idea, it was never going to replace Python/C++. Eventually Lattner lost interest and the project died.
reply