lite_rt_for_flutter 0.0.1

lite_rt_for_flutter: ^0.0.1 copied to clipboard

IMPORTANT, READ THIS BEFORE USING THIS PLUGIN!

• This is NOT an official implementation, Google's official plugin is here
  • For proper support please use that plugin
• I created this plugin to use in my own applications as I encountered various difficulties with the official plugin.
  • Therefore, the main use case is that it works in my own applications, but it may be useful for others, which is why I opened sourced it.

LiteRT (TF Lite) for Flutter #

This plugin provides bindings for LiteRT (formerly TF lite) for Flutter. Try it out in the browser!

The only real difference between this (except some convenience methods) and the standalone dart package is that LiteRT binaries can be included using the additional libs packages:

✅ - supported ; 🚧 - working on it ; 📋 planned ; ⛔ - not applicable ; ❌ - not supported by LiteRT ; ℹ️ - click to learn more

Setup #

In the dependency section of pubspec.yaml file, add lite_rt_for_flutter: <your version>

dependencies:
  # The binding between Flutter/Dart and native code
  lite_rt_for_flutter:
    version: <THE_VERSION_YOU_WANT>

  # Additionally, you need the base LiteRT libraries (includes XNNPack delegate)
  lite_rt_for_flutter_libs:
    version: <THE_VERSION_YOU_WANT>
  # optionally you can set specific versions per platform
  lite_rt_for_flutter_libs_{android|macos|...}:
    version: <THE_VERSION_YOU_WANT>

Finally, run some ML on-device!

import 'package:liter_rt_for_flutter/liter_rt_for_flutter.dart';

// IMPORTANT: initialize the plugin
initLiteRTFlutter()

// Load a model from asset
final interpreter = await FlutterInterpreter.fromAsset('asset/path/to/your/model.tflite');

You can learn more about how to use this library here:

• the dart package.
• For som complete examples, see the examples folder

What about delegates? #

Additional delegate libraries can be included for supported platforms. See the above table for which delegate supports which platform.

dependencies:
  # Include the CoreML delegate
  lite_rt_for_flutter_libs_coreml:
    version: <THE_VERSION_YOU_WANT>
  # Include the flex delegate
  lite_rt_for_flutter_libs_flex:
    version: <THE_VERSION_YOU_WANT>
  # Include the gpu delegate
  lite_rt_for_flutter_libs_gpu:
    version: <THE_VERSION_YOU_WANT>

Web support #

This library does NOT directly support web. However, using the following hacky approach, web support can be achieved by using the tflit web package. Under the hood it uses tfjs' tf lite support using webassembly.

Binaries can be included using

dependencies:
  # includes xnnpack delegate
  lite_rt_for_flutter_libs_web:
    version: 0.0.1-alpha10

Now you need to define inference based on native / web, for an example see [example/all_platforms_classification].

I want to use my own libs #

Sure thing! See the dart package to use any library.