The model was trained on d2go. The architecture can be found at Mask-RCNN-FBNetV3A-dsmask.
The demo app is located in maskrcnn/ directory. To run the app, simply do pod install. For the custom ops, we're using eigen(3.3.9) for vectorization. The code was written in C++.
Make sure you have PyTorch installed on your machine.
> import torch
> torch.__version__ # '1.9.0'
Run the converter
> python converter.py
CoreML-MaskRCNN has a MIT-style license, as found in the LICENSE file.