I will try or ask my friend to do.
Amit Shekhar

That would be awesome mate! We’ve been trying to withot success, but we did come across some good resources. “For larger models — and especially for deep learning — you want to use the GPU instead. And on iOS that means using Metal. You’ll still do the training with TensorFlow on your Mac — or for big models on a Linux box with a fat GPU or even in the cloud — but the inference code that runs on iOS uses Metal instead of the TensorFlow library.” For ex he provides the steps in order to use Metal http://machinethink.net/blog/tensorflow-on-ios/ 
Also this guy manage to do it, presenting the caveats too: http://blog.stratospark.com/creating-a-deep-learning-ios-app-with-keras-and-tensorflow.html

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.