Microsoft SEAL is an easy-to-use open-source (MIT licensed) homomorphic encryption library developed by the Cryptography Research group at Microsoft.

pybind11 is a lightweight header-only library that exposes C++ types in Python and vice versa, mainly to create Python bindings of existing C++ code.

• This implementation contains noise-flooding to protect the circuit privacy of Homomorphic Encryption
• This repo is dedicated to GForce: GPU-Friendly Oblivious and Rapid Neural Network Inference, but it is also welcome for other uses.

• Build
• Tests
• About
• Contributing

• Environment

  CMake (>= 3.10), GNU G++ (>= 6.0) or Clang++ (>= 5.0), Python (>=3.6.8)

  sudo apt-get update && sudo apt-get install -y g++ make git gcc-8 g++-8

  git clone https://github.com/Lucieno/SEAL-Python.git

  We highly recommend you to install SEAL-Python in a Anaconda virtual environment.

• SEAL 3.3.2

  cd SEAL/native/src
  cmake .
  make
  sudo make install

• pybind11

  conda install pybind11
  

• SEAL-Python

  cd src
  pip install -r requirements.txt
  
  export CC='/usr/bin/gcc-8'
  export CXX='/usr/bin/g++-8'
  python setup.py install
  

• Others

  If you clone a new SEAL lib from the Github, do not forget add a function set_scale in seal/ciphertext.h line 632, like this:

  /**
  Set the scale.
  */
  inline void set_scale(double scale)
  {
    scale_ = scale;
  }

  The setuptools will build a dynamic Link Library in project folder, named like seal.cpython-36m-x86_64-linux-gnu.so.

  You can use the CMake to build it, I have already written the CMakelists.txt. Make sure the SEAL and the pybind11 is correctly installed.

  The path is very important, please check it before you do anything.

cd tests

python [example_name].py

• The .so file need in the same folder, or you had make install it already.