lite_rt_for_dart 0.0.2 
lite_rt_for_dart: ^0.0.2 copied to clipboard
Plugin to run tf lite models in dart apps.
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 dart #
This plugin provides bindings for LiteRT (formerly TF lite) for standalone dart. If you want to use this in Flutter, consider using the Flutter package. Try it out in the browser!
Why another plugin? #
This fork of the official plugin tries to improve upon the official implementation in a few key areas:
- Dart standalone support
- Improved binary handling
- Up-to-date binaries
- Easily switch between binaries
- Improved platform support
- Linux arm 64
- Windows arm 64
- Web support is planned
New features from the official implementation will also be integrated in this fork.
Example #
Import and initialize the library #
In the dependency section of pubspec.yaml file, add lite_rt_for_dart: <your version>
import 'package:lite_rt_for_dart/tflite_flutter.dart';
// IMPORTANT: initialize the plugin
initLiteRT("path/to/your/lib_tflite")
// Load a model from file
final interpreter = await Interpreter.fromFile('path/to/your/model.tflite');
Refer to the documentation for info on creating interpreter from a buffer or a file.
Performing inference #
For single input and output
Use void run(Object input, Object output).
// For ex: if input tensor shape [1,5] and type is float32
var input = [[1.23, 6.54, 7.81, 3.21, 2.22]];
// if output tensor shape [1,2] and type is float32
var output = List.filled(1*2, 0).reshape([1,2]);
// inference
interpreter.run(input, output);
// print the output
print(output);
For multiple inputs and outputs
Use void runForMultipleInputs(List<Object> inputs, Map<int, Object> outputs).
var input0 = [1.23];
var input1 = [2.43];
// input: List<Object>
var inputs = [input0, input1, input0, input1];
var output0 = List<double>.filled(1, 0);
var output1 = List<double>.filled(1, 0);
// output: Map<int, Object>
var outputs = {0: output0, 1: output1};
// inference
interpreter.runForMultipleInputs(inputs, outputs);
// print outputs
print(outputs)
Closing the interpreter
interpreter.close();
Asynchronous Inference with IsolateInterpreter #
To utilize asynchronous inference, first create your Interpreter and then wrap it with IsolateInterpreter.
final interpreter = await Interpreter.fromAsset('assets/your_model.tflite');
final isolateInterpreter =
await IsolateInterpreter.create(address: interpreter.address);
Both run and runForMultipleInputs methods of isolateInterpreter are asynchronous:
await isolateInterpreter.run(input, output);
await isolateInterpreter.runForMultipleInputs(inputs, outputs);
By using IsolateInterpreter, the inference runs in a separate isolate. This ensures that the main isolate, responsible for UI tasks, remains unblocked and responsive.
Development #
Metadata
Publisher
unverified uploader
Weekly Downloads
Metadata
Plugin to run tf lite models in dart apps.
Repository (GitHub)
View/report issues
Documentation
License
Apache-2.0 (license)
