Export object detection models to Android

Hi there, I’ve been using GluonCV for object detection and action recognition, and I was trying to figure out how to export these models to Android (as for object detection, I re-trained the algorithm on my own dataset, so I’d like to export that).

I’m really new to the topic, so I was looking for a place to start. Should I use an external tool (like BeeWare or Kivy), or can I do everything by native mxnet/GluonCV tools?
I found this discussion, but as you can see the first link in the answer is not working anymore, and instructions in the amalgamation repository are not particularly clear. I tried to clone the repo and run make from amalgamation subfolder, but I’m repeatedly getting this error:

Makefile:80: recipe for target 'nnvm.d' failed
make[3]: *** [nnvm.d] Error 1
make[3]: Leaving directory '/home/lews/PycharmProjects/incubator-mxnet/amalgamation'
cp: cannot stat 'nnvm.d': No such file or directory
cat: nnvm.cc: No such file or directory
mv: cannot move 'temp' to '../../../../amalgamation/nnvm.cc': No such file or directory

Any indication would be very appreciated, thanks!