LibML is a library for loading, configuring, and running machine learning models in production. It provides a simple high-level API for C++ applications. LibML uses TensorFlow with XNNPACK acceleration for low latency inference.
./configure.sh
cd build
sudo make -j$(nproc) install
BinaryClassifier classifier;
if(!classifier.buildFromFile(model_path))
return 1;
float output = 0.0;
if(!classifier.run(input, input_size, output))
return 1;
std::cout << "output: " << output << "%\n";
- CMake
- C++ Compiler