Project Structure
CoreLocker
├── README.md
├── corelock.py
├── extractor.py
└── models
├── __init__.py
├── channel_selection.py
├── densenet.py
└── preresnet.py
To split the target model:
python corelock.py split --model <path_to_model> --ratio <key_ratio> --arch <architecture> --save ./test_output
Enable --config to observe the configeration post extraction.
To patch the model:
python corelock.py recover --model test_output/keyed_model.pth.tar --key test_output/key.pth.tar --arch <architecture> --save <save_path>