Full disclaimer, I'm not an ML expert by any stretch of the imagination and that probably shows in this blog post. But I have recently needed to run some simple models in Go that I've trained in TensorFlow and this approach has worked great for me so far, so I wanted to share it :).