Welcome to Intel® nGraph™, an open source C++ library and compiler. This project enables modern compute platforms to run and train Deep Neural Network models. It is framework-neutral and supports a variety of backends used by Deep Learning frameworks.
For this early release, we’ve provided Framework Integration Guides to compile and run MXNet* and TensorFlow*-based projects. If you already have a trained model, see our section on How to Import a model and start working with the nGraph APIs.
The library code is under active development as we’re continually adding support for more ops, more frameworks, and more backends.
The nGraph++ library translates a framework’s representation of computations
into an Intermediate Representation that promotes computational
efficiency on target hardware. Initially-supported backends include Intel
Architecture CPUs (
CPU), the Intel® Nervana Neural Network Processor™ (NNP),
and NVIDIA* GPUs. Currently-supported compiler optimizations include efficient
memory management and data layout abstraction.
Further project details can be found on our About page.