In order to convert prelearned models such as Caffe, Keras, e.t.c. into .mlmodel file extension you have to do so with convertor that you can install via PIP.
PIP - package management system that helps you install packages
0) Install python if needed
brew install python
1) Save this file
2) cd to this file and install it with
sudo python get-pip.py
3) Finally, install coreML tools by Apple
pip install -U coremltools
Hint: -U means latest version
4) Create "convert-script.py" in the directory the model placed. It should be looking something like that
import coremltools
caffe_model = ('oxford102.caffemodel', 'deploy.prototxt')
labels = 'flower-labels.txt'
coreml_model = coremltools.converters.caffe.convert(caffe_model, class_labels=labels, image_input_names='data')
coreml_model.save('FlowerClassifier.mlmodel')
5) Run the script
pip install -U coremltools
The Converted model file should be in the directory