Samsung/ONE

On-device Neural Engine

What it solves

It provides a high-performance framework for running neural network inference directly on devices, avoiding the need for cloud-based processing. It solves the problem of executing AI models across different hardware accelerators like CPUs, GPUs, DSPs, and NPUs on Linux-based operating systems.

How it works

ONE uses a compiler toolchain to convert neural network models trained in frameworks like TensorFlow or PyTorch into a unified format. This unified format is then executed by a dedicated runtime that supports various hardware backends on platforms such as Ubuntu, Android, and Tizen.

Who it’s for

It is designed for developers who need to deploy AI models to on-device hardware and optimize performance across different processors.

Highlights

  • High-performance on-device inference.
  • Support for multiple hardware accelerators (CPU, GPU, DSP, NPU).
  • Support for Linux-kernel based OS platforms (Ubuntu, Tizen, Android).
  • Unified runtime for models from TensorFlow and PyTorch.

Related

  • Project
    microsoft/onnxruntime

    A cross-platform machine learning accelerator that optimizes inference and training for models from various frameworks like PyTorch and TensorFlow.

  • Project
    uxlfoundation/oneDNN

    A cross-platform performance library providing optimized building blocks for deep learning applications to improve performance across various CPUs and GPUs.

  • Project
    sonos/tract

    A neural-network inference engine that optimizes and deploys AI models across embedded ARM, GPUs, and WebAssembly with a focus on a tiny runtime footprint.

  • Project
    tracel-ai/burn

    Burn is a unified tensor library and deep learning framework written in Rust that allows the same code to be used for both training and inference across diverse hardware.

  • Project
    nndeploy/nndeploy

    A high-performance AI deployment framework that uses visual workflows to deploy AI models across desktop, mobile, and edge devices.