Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Neural Network Exchange Format
open-in-new
<h2 id="history">History</h2> <p>NNEF was proposed in 2015 by member companies of the Khronos Group as a device and implementation independent transfer format capable of describing any artificial neural net in terms of its structure, operations and data. </p><p>The first version of the standard was launched in provisional form in December 2017, and was ratified as an official Khronos standard in August 2018. </p> <h2 id="objectives">Objectives</h2> <p>The goal of NNEF is to enable data scientists and engineers to easily transfer trained networks from their chosen training framework into a wide variety of inference engines. NNEF encapsulates a complete description of the structure, operations and parameters of a trained neural network, independent of the training tools used to produce it and the inference engine used to execute it. </p> <h2 id="governance-and-availability">Governance and Availability</h2> <p>NNEF is maintained by the Khronos Group under its Open Governance Principles<a class="footnote-ref" id="fnref:3" href="#fn:3"><sup>3</sup></a> as follows: </p> <ul><li>Any company is invited and able to join Khronos to contribute to and influence the development of its specifications;</li> <li>Finalized specifications are publicly and freely distributed at zero cost from the Khronos web-site;</li> <li>Any company can implement a Khronos specification and participating implementers can obtain a trademark license for conformant implementations and pay zero royalties to Khronos participants; and</li> <li>Developers may freely use implementations of Khronos specifications.</li></ul> <p>The NNEF specification is available on the <a href="https://www.khronos.org/registry/NNEF/">Khronos NNEF registry</a> and tools are available on <a href="https://github.com/KhronosGroup/NNEF-Tools">Github</a> </p> <h2 id="versions">Versions</h2> <ul><li>NNEF 1.0 Provisional, Released 20 December 2017.<a class="footnote-ref" id="fnref:4" href="#fn:4"><sup>4</sup></a></li> <li>NNEF 1.0, Released 13 August 2018<a class="footnote-ref" id="fnref:5" href="#fn:5"><sup>5</sup></a> <ul><li>NNEF 1.0.1, Released 10 May 2019</li> <li>NNEF 1.0.2, Released 13 July 2019<a class="footnote-ref" id="fnref:6" href="#fn:6"><sup>6</sup></a></li></ul></li></ul> <h2 id="industry-participation">Industry Participation</h2> <p>The following Khronos members have participated in the NNEF working group: </p> <ul><li>AIMotive.</li> <li>Advanced Micro Devices.</li> <li>Arm Holdings, Ltd.</li> <li>Axell</li> <li>Axis Communications.</li> <li>Cadence</li> <li>Ceva</li> <li>Codeplay</li> <li>Digital Media Professionals</li> <li>ETRI</li> <li>Huawei</li> <li>Intel Corp.</li> <li>Imagination technologies</li> <li>LG</li> <li>Los Alamos National Lab</li> <li>LunarG</li> <li>Mediatek</li> <li>Mentor Graphics</li> <li>NXP</li> <li>On Semiconductor</li> <li>Qualcomm</li> <li>The Qt Company</li> <li>Renesas</li> <li>Samsung</li> <li>Silicon Studio</li> <li>Socionext</li> <li>Sony</li> <li>Synopsys</li> <li>Texas Instruments</li> <li>Think Silicon</li> <li>Verisilicon</li> <li>Xilinx</li></ul> <h2 id="tools">Tools</h2> <p>The <a href="https://github.com/KhronosGroup/NNEF-Tools">NNEF tools project</a> on GitHub contains the following open source tools: </p> <ul><li>File format Parser</li> <li>Bidirectional converters between NNEF and ONNX, Caffe, Caffe2, TensorFlow (python), TensorFlow (protobuf)</li> <li>Model zoo: reference collection of models converted to NNEF</li></ul> <h2 id="see-also">See also</h2> <ul><li><a href="/facts/Open_Neural_Network_Exchange/mjR9T7hd">Open Neural Network Exchange</a></li></ul> <h2 id="references">References</h2> <ol> <li id="fn:1"><p>"NNEF - Neural Network Exchange Format (NNEF)". The Khronos Group. 2016-10-04. Retrieved 2019-02-07. <a href="https://www.khronos.org/nnef/" target="_blank">https://www.khronos.org/nnef/</a> <a href="#fnref:1" class="footnote-back-ref">↩</a></p></li> <li id="fn:2"><p>Seo, B.; Shin, M.; Mo, Y. J.; Kim, J. (January 2018). "Top-down parsing for Neural Network Exchange Format (NNEF) in TensorFlow-based deep learning computation". 2018 International Conference on Information Networking (ICOIN). pp. 522–524. doi:10.1109/ICOIN.2018.8343173. ISBN 978-1-5386-2290-2. S2CID 5053900. <a href="978-1-5386-2290-2" target="_blank">978-1-5386-2290-2</a> <a href="#fnref:2" class="footnote-back-ref">↩</a></p></li> <li id="fn:3"><p>Khronos IP Framework <a href="https://www.khronos.org/members/ip-framework/" target="_blank">https://www.khronos.org/members/ip-framework/</a> <a href="#fnref:3" class="footnote-back-ref">↩</a></p></li> <li id="fn:4"><p>v1.0p Khronos PR <a href="https://www.khronos.org/news/press/khronos-group-releases-nnef-1.0-standard-for-neural-network-exchange" target="_blank">https://www.khronos.org/news/press/khronos-group-releases-nnef-1.0-standard-for-neural-network-exchange</a> <a href="#fnref:4" class="footnote-back-ref">↩</a></p></li> <li id="fn:5"><p>"The Khronos Group launches new standard for deploying trained neural networks". SD Times. 2018-08-13. Retrieved 2019-02-11. <a href="https://sdtimes.com/ai/the-khronos-group-launches-new-standard-for-deploying-trained-neural-networks/" target="_blank">https://sdtimes.com/ai/the-khronos-group-launches-new-standard-for-deploying-trained-neural-networks/</a> <a href="#fnref:5" class="footnote-back-ref">↩</a></p></li> <li id="fn:6"><p>"Khronos NNEF Registry - The Khronos Group Inc". www.khronos.org. Retrieved 2019-08-15. <a href="https://www.khronos.org/registry/NNEF/" target="_blank">https://www.khronos.org/registry/NNEF/</a> <a href="#fnref:6" class="footnote-back-ref">↩</a></p></li> </ol>