torch_mobile 0.0.2

torch_mobile #

A Flutter plugin for inference of Pytorch models.

Note: This plugin is still under development, only image classification models are supported for the moment.

Installation #

First, add torch_mobile as a dependency in your pubspec.yaml file.

iOS #

Not implemented yet

Android #

No configuration required - the plugin should work out of the box.

Usage #

Create a assets folder with pytorch model and labels file and model file in it. Modify pubspec.yaml accordingly.

  assets:
   - assets/model.pt
   - assets/labels.txt

Import the library

import 'package:torch_mobile/torch_mobile.dart';

Load model and labels

TorchMobile.loadModel(model: 'assets/model.pt', labels: 'assets/labels.txt');

Get prediction for image

String prediction = await TorchMobile.getPrediction(image, maxWidth: 400, maxHeight: 400);

0.0.2 #

  • Update path lib dependency

0.0.1 #

  • Initial Release

example/README.md

torch_mobile_example #

Demonstrates how to use the torch_mobile plugin.

Example #

import 'dart:io';

import 'package:flutter/material.dart';
import 'package:flutter/services.dart';
import 'dart:async';
import 'package:image_picker/image_picker.dart';
import 'package:torch_mobile/torch_mobile.dart';

void main() {
  SystemChrome.setPreferredOrientations(
      [DeviceOrientation.portraitUp, DeviceOrientation.portraitDown]).then((_) {
    runApp(SampleApp());
  });
}

class SampleApp extends StatefulWidget {
  @override
  _SampleAppState createState() => _SampleAppState();
}

class _SampleAppState extends State<SampleApp> {
  String _prediction = '';
  File _image;

  @override
  void initState() {
    super.initState();
    try {
      TorchMobile.loadModel(
          model: 'assets/model.pt', labels: 'assets/labels.txt');
    } on PlatformException {}
  }

  Future getImage() async {
    var image = await ImagePicker.pickImage(
        source: ImageSource.camera, maxWidth: 400.0, maxHeight: 400.0);
    await makePrediction(image);
    setState(() {
      _image = image;
    });
  }

  // Platform messages are asynchronous, so we initialize in an async method.
  Future<void> makePrediction(File file) async {
    String prediction;
    try {
      prediction =
          await TorchMobile.getPrediction(file, maxWidth: 400, maxHeight: 400);
    } on PlatformException {
      prediction = 'Failed to get prediction.';
    }
    if (!mounted) return;

    setState(() {
      _prediction = prediction;
    });
  }

  @override
  Widget build(BuildContext context) {
    return MaterialApp(
      home: Scaffold(
        appBar: AppBar(
          title: const Text('Torch mobile image prediction app'),
        ),
        body: Column(
          mainAxisAlignment: MainAxisAlignment.spaceAround,
          children: <Widget>[
            if (_image != null) Center(child: Image.file(_image)),
            if (_prediction != null)
              Padding(
                padding: const EdgeInsets.all(8.0),
                child: Text(
                  '$_prediction\n',
                  textAlign: TextAlign.center,
                ),
              ),
          ],
        ),
        floatingActionButton: FloatingActionButton(
          onPressed: getImage,
          tooltip: 'Pick Image',
          child: Icon(Icons.add_a_photo),
        ),
      ),
    );
  }
}

Use this package as a library

1. Depend on it

Add this to your package's pubspec.yaml file:


dependencies:
  torch_mobile: ^0.0.2

2. Install it

You can install packages from the command line:

with Flutter:


$ flutter pub get

Alternatively, your editor might support flutter pub get. Check the docs for your editor to learn more.

3. Import it

Now in your Dart code, you can use:


import 'package:torch_mobile/torch_mobile.dart';
  
Popularity:
Describes how popular the package is relative to other packages. [more]
45
Health:
Code health derived from static analysis. [more]
100
Maintenance:
Reflects how tidy and up-to-date the package is. [more]
90
Overall:
Weighted score of the above. [more]
70
Learn more about scoring.

We analyzed this package on Feb 13, 2020, and provided a score, details, and suggestions below. Analysis was completed with status completed using:

  • Dart: 2.7.1
  • pana: 0.13.5
  • Flutter: 1.12.13+hotfix.7

Maintenance suggestions

Package is pre-v0.1 release. (-10 points)

While nothing is inherently wrong with versions of 0.0.*, it might mean that the author is still experimenting with the general direction of the API.

Dependencies

Package Constraint Resolved Available
Direct dependencies
Dart SDK >=2.2.0 <3.0.0
flutter 0.0.0
path ^1.6.4 1.6.4
path_provider ^1.3.1 1.6.0
Transitive dependencies
collection 1.14.11 1.14.12
meta 1.1.8
platform 2.2.1
sky_engine 0.0.99
typed_data 1.1.6
vector_math 2.0.8
Dev dependencies
flutter_test