https://github.com/kashifmin/flutter_tensorflow_lite

A Flutter plugin to access TensorFlow Lite apis.
https://github.com/kashifmin/flutter_tensorflow_lite

android dart dartlang flutter flutter-plugin ios kotlin tensorflow

Last synced: 20 days ago
JSON representation

A Flutter plugin to access TensorFlow Lite apis.

• Host: GitHub
• URL: https://github.com/kashifmin/flutter_tensorflow_lite
• Owner: kashifmin
• License: bsd-3-clause
• Created: 2018-04-23T12:50:51.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-07-28T10:32:11.000Z (over 6 years ago)
• Last Synced: 2024-11-16T03:39:11.469Z (about 2 months ago)
• Topics: android, dart, dartlang, flutter, flutter-plugin, ios, kotlin, tensorflow
• Language: Dart
• Homepage: https://pub.dartlang.org/packages/tensorflow_lite
• Size: 7.9 MB
• Stars: 74
• Watchers: 6
• Forks: 7
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# tensorflow_lite

[![pub package](https://img.shields.io/pub/v/tensorflow_lite.svg)](https://pub.dartlang.org/packages/tensorflow_lite)

A Flutter plugin to access TensorFlow Lite apis.

[TensorFlow Lite](https://www.tensorflow.org/mobile/tflite/) is TensorFlow’s lightweight solution for mobile and embedded devices.

With TensorFlow Lite you can deploy machine learning models on phones in your Android/iOS app.

## Usage

Add `tensorflow_lite` to your [pubspec.yaml](https://flutter.io/platform-plugins/)

Copy your models to an asset dir like `assets/mobilenet_quant_v1_224.tflite`

And add it to your pubspec.yaml

```yaml

   assets:

     - assets/mobilenet_quant_v1_224.tflite

```

Import tensorflow_lite in your app

```dart

import 'package:tensorflow_lite/tensorflow_lite.dart';

```

Create a new Interpreter instance based on your tflite model file

```dart

Interpreter model = await Interpreter.createInstance(modelFilePath: modelPath);

```

Pass some bytes to the model to get the output

```dart

dynamic result = await _interpreter.run(imageToByteList(image), new Uint8List(_labelList.length));

```

# Image Classification example

`tensorflow_lite` also includes a wrapper for image classification models which can be easily loaded

without much of boilerplate code.

```dart

Future loadRecognitions() async {

    var classifier = await TFLiteImageClassifier.createInstance(

      assets: rootBundle,

      modelPath: "assets/mobilenet_quant_v1_224.tflite",

      labelPath: "assets/labels.txt",

      inputSize: 224,

    );

    print('Classifier ready');

    var imageBytes = (await rootBundle.load("assets/cat500.png")).buffer;

    img.Image image = img.decodePng(imageBytes.asUint8List());

    image = img.copyResize(image, 224, 224);

    _recognitions = await classifier.recognizeImage(image);

    setState(() {});

    await classifier.close();

  }

```

Please check the example for full usage.

## Note

- Works only on Android

- Tested only on image classification

## Contributing

I am new to Flutter and I haven't worked on iOS yet.

So if you are an iOS developer, i'd be glad to receive some contribution.

Just send a PR or open up an issue!