PyTorch -> TorchScript -> CoreML -> ANE Execution
https://machinelearning.apple.com/research/neural-engine-transformers
https://github.com/apple/ml-ane-transformers
Need to see the limits imposed by Apple on compute capabilities
https://www.youtube.com/watch?v=JAyw7OAcXDE&t=8490s
Shoutout George for maybe even more finesse
https://github.com/geohot/tinygrad/tree/master/accel/ane