I had a Tensorflow model which I was ready to move to the next stage of production but had difficulties with some of the standard tools. CoreML and Serving didn’t seem to work for my use-case. I believe they’d work if I had structured model inference with the TF graph rather than procedurally.
How would you put a trained ML model into production? Any recommended learning resource?