Getting Started

Installation

Install from source:

pip install .

Or with Pixi:

pixi install
pixi run install

Or build with CMake (C++ only):

cmake -S . -B build -DONNX_LIGHT_CPU_BUILD_TESTS=ON \
      -DONNX_LIGHT_CPU_BUILD_PYTHON=OFF
cmake --build build

Quick Start (C++)

Include the public header and call one of the kernel functions; the best available SIMD path is selected automatically at runtime:

#include <onnx_light_cpu/cpu_kernels.h>

int main() {
    float input[] = {-1.0f, 2.0f, -3.0f, 4.0f};
    float output[4];
    onnx_light_cpu::AbsFloat32(input, output, 4);
    // output = {1.0f, 2.0f, 3.0f, 4.0f}
}

Quick Start (Python)

import numpy as np
from onnx_light_cpu.onnx_py._cpukernels import abs, detect_simd_level

# Check what SIMD level is available:
# 0=None, 1=SSE2, 2=AVX, 3=AVX2, 4=AVX512
print("SIMD level:", detect_simd_level())

inp = np.array([-1.0, 2.0, -3.0, 4.0], dtype=np.float32)
out = abs(inp)  # dispatches on dtype, returns a new array like numpy.abs
print(out)  # [1. 2. 3. 4.]