Contents

• Initial setup if using a pre-built package
  • Android
  • iOS
  • Install ONNX Runtime python package
• Initial setup if performing a custom build
  • Clone ONNX Runtime repository
  • Install ONNX Runtime python package

Initial setup if using a pre-built package

Android

Java/Kotlin

In your Android Studio Project, make the following changes to:

1. build.gradle (Project):

    repositories {
        mavenCentral()
    }
   

2. build.gradle (Module):

    dependencies {
        compile 'com.microsoft.onnxruntime:onnxruntime-mobile:1.8.0'
    }
   

C/C++

Download the onnxruntime-mobile AAR hosted at MavenCentral, change the file extension from .aar to .zip, and unzip it. Include the header files from the headers folder, and the relevant libonnxruntime.so dynamic library from the jni folder in your NDK project.

iOS

In your CocoaPods Podfile, add the onnxruntime-mobile or onnxruntime-mobile-objc pod depending on which API you wish to use.

Run pod install.

C/C++

  use_frameworks!

  pod 'onnxruntime-mobile'

Objective-C

  use_frameworks!

  pod 'onnxruntime-mobile-objc'

Install ONNX Runtime python package

Install the onnxruntime python package from https://pypi.org/project/onnxruntime/ in order to convert models from ONNX format to the internal ORT format. Version v1.8 or higher is required.

• pip install onnxruntime will install the latest release

Initial setup if performing a custom build

Clone ONNX Runtime repository

Use git to clone the ONNX Runtime repository

• git clone --recursive https://github.com/Microsoft/onnxruntime
  • this will create an ‘onnxruntime’ directory with the repository contents
• See the Build for inferencing documentation for further details on supported environments. Ignore the build instructions on that page as they are for a full build and we will cover the mobile build instructions here.

Select the branch you wish to use. The latest release is recommended.

• git checkout <branch>
  • e.g. git checkout rel-1.8.0

It is suggested you do not use the unreleased ‘master’ branch unless there is a specific new feature you require.

The directory the ONNX Runtime repository was cloned into is referred to as <ONNX Runtime repository root> in this documentation.

Install ONNX Runtime python package

Install the onnxruntime python package from https://pypi.org/project/onnxruntime/ in order to convert models from ONNX format to the internal ORT format. Version 1.5.3 or higher is required.

• pip install onnxruntime will install the latest release

You must match the python package version to the branch of the ONNX Runtime repository you checked out

• e.g. if you wanted to use the 1.7 release
  • git checkout rel-1.7.2 in your local git repository
  • pip install onnxruntime==1.7.2

If you are using the master branch in the git repository you should use the nightly ONNX Runtime python package

• pip install -U -i https://test.pypi.org/simple/ ort-nightly

Next: Converting ONNX models to ORT format