English | 简体中文
Model Crypto is a versatile Python and C++ library for encrypting and decrypting machine learning models. It's built on Crypto++ and offers support for various deep learning frameworks such as TensorRT, PyTorch, Paddle, and more. Use Model Crypto to safeguard your AI models with robust encryption techniques.
- Windows: Microsoft Visual Studio (Tested with Visual Studio 2019, 2022)
- XMake (Latest version recommended)
Compile your C++ code with the following commands, specifying the target platform
and architecture
:
xmake f -p {platform} -a {architecture} -m release
xmake -w
After compilation, you will find the C++ library in the project's root directory, including the include
and lib
directories. Additionally, the Python library required for building is located in .pyd
format within the python\model_crypto\libs
directory.
To install the Python library, use the following command:
pip install .
Model Crypto is provided under the MIT.