https://github.com/ZHKKKe/MODNet
create a directory named "pretrained"
Pre-trained model can be downloaded using following command:
$ gdown --id 1mcr7ALciuAsHCpLnrtG_eop5-EYhbCmz -O pretrained/modnet_photographic_portrait_matting.ckpt
or use the following link to download directly from google drive:
https://drive.google.com/drive/folders/1umYmlCulvIFNaqPjwod1SayFmSRHziyR?usp=sharing
$ pip install -r requirements.txt
Note: Demo Tested with torch==1.1.0 and torchvision==0.3.0
python inference.py --input example.jpg