eIQ EAR1 for i.MX Release Announcement

Document created by Ragan_Dunham Employee on Mar 26, 2019Last modified by Karina Valencia Aguilar on Apr 17, 2019
Version 7Show Document
  • View in full screen mode

Today NXP is announcing the immediate availability of eIQ EAR1 for i.MX, a pre-release alpha version of ML software components for NXP i.MX processors. The NXP eIQ™ Machine Learning Software Development Environment represents the key ingredients to deploy a wide range of ML algorithms; it includes inference engines, NN compilers, and hardware abstraction layers. This suite of tools is under active development and is rapidly expanding to further meet your needs to develop and deploy full-featured ML solutions.

 

The attached eIQ User's Guide describes how to build the main libraries for AI/ML, including support for Arm NN (for TensorFlow, TensorFlow Lite, Caffe, and ONNX models) and OpenCV (for TensorFlow and Caffe models)-- using an NXP standard Yocto Linux release. Also included are instructions for building and installing a few sample applications to help get you up and running more quickly.

 

 

Support

This software is provided as-is, without warranty and without support.

This software is alpha quality -- there might be bugs! Known limitations are documented where applicable.

 

Hardware requirements

It is known to run on the NXP evaluation boards for the i.MX 8 Series Applications Processors: i.MX 8QM MEK, i.MX 8QXP MEK, i.MX 8M EVK, and i.MX 8M Mini EVK. No other platforms have been specifically tested.

 

Documentation

Attached please find the eIQ User Guide Rev. v1_EAR1 , 03/2019.

 

Source code

To download the source code please visit the README [1] linked below.

The meta-imx-machinelearning [2] provides some machine learning related recipes such as OpenCV (3.4.1) and ArmNN(18.08) on top of the current standard NXP Yocto Linux release L4.14.78-1.0.0 GA.

 

[1] https://source.codeaurora.org/external/imx/imx-manifest/tree/README?h=imx-linux-sumo

[2] https://source.codeaurora.org/external/imx/meta-imx-machinelearning/tree/?h=sumo

Attachments

Outcomes