Skip to content

vlmoon99/ai_model_land

BranchesTags

Repository files navigation

AIModelLand

About AIModelLand
Installing
How to use
Running Examples
Contributing

About AIModelLand

AIModelLand is a library that provides developers with a simple API for interacting with various AI providers. It allows developers to easily integrate popular AI providers into their projects. Using this library, developers don't have to waste time connecting a new provider to a project and can easily use all the functionality that the provider and AI model provide without any difficulty.

Installing

  1. Add this code to the index.html for Web projects:

    <head>
  <!-- other head elements -->
  <script type="application/javascript" src="flutter.js" defer></script>
  <script
    type="application/javascript"
    src="/assets/packages/flutter_inappwebview_web/assets/web/web_support.js"
    defer
  ></script>
  <script
    type="module"
    src="/assets/packages/ai_model_land/assets/onnx/index.js"
    defer
  ></script>
</head>
<script>
  window.addEventListener("load", async function (ev) {
    {
      {
        flutter_js;
      }
    }
    {
      {
        flutter_build_config;
      }
    }

    _flutter.loader.load({
      serviceWorker: {
        serviceWorkerVersion: "{{flutter_service_worker_version}}",
      },
      onEntrypointLoaded: function (engineInitializer) {
        engineInitializer.initializeEngine().then(function (appRunner) {
          appRunner.runApp();
        });
      },
    });
  });
</script>

  2. Set Android: minSdkVersion >= 26, compileSdk >= 34, AGP version >= 7.3.0 in the android/app/build.gradle file and add to the AndroidManifest.xml the usesCleartextTraffic parameter and Internet permission:

    <manifest> ...
  <application> ...
    android:usesCleartextTraffic="true"
  </application>
  <uses-permission android:name="android.permission.INTERNET"/>
</manifest>

  3. Set Flutter version >= 3.24.3 stable.
    Example code for updating to the latest Flutter stable version:

    flutter channel stable
flutter upgrade

  4. Initialize the AIModelLand library after WidgetsFlutterBinding.ensureInitialized() in the main function:

    void main() {
  WidgetsFlutterBinding.ensureInitialized();
  initAIModelLandLib();
  ...
  runApp(MyApp());
}

How to use

With the library, you can easily manage the most popular AI providers without any problems. To initialize AIModelLand, we can use the default constructor and the named constructor defaultInstance().

Example:

final aiModelLand = AiModelLandLib.defaultInstance();

The next step is to create base information about the model by using the BaseModel class.

In this class, we have the following parameters:

  • sourceType with enum ModelSourceType – choose where your model is located. If you choose network, the file will be downloaded to your device.
  • source – in this parameter, input the file location. It can be a file explorer path or an assets path. If you choose network in sourceType, input the URL to the model. If in the future you will load the model to the provider using uint8List format, you can input an empty string.
  • nameFile – input the existing file name, or if you download the model, this will be the future file name.
  • format with enum ModelFormat – choose the provider that you want to use.

Then you can reuse this BaseModel in all functions for each AI provider.

You can check how to use all functions for each AI provider by the following links:

Example of using the AIModelLand library:

//create base model
BaseModel baseModel = BaseModel(
    source: "assets/onnx/image_classification/mobilenetv2-7.onnx",
    nameFile: "photo-detection-classification",
    format: ModelFormat.onnx,
    sourceType: ModelSourceType.local);

//load model to provider
bool isLoadModel = await aiModelLand.loadModel(
        request: OnnxRequestModel(
          loadModelWay: LoadModelWay.fromAssets,
          onnxBackend: ONNXBackend.cpu ),
        baseModel: baseModel);

//run model (get prediction)
OnnxResponsModel predict = await aiModelLand.runTaskOnTheModel(
        request: OnnxRequestModel(dataMulti: [
          inputBytes
        ], shape: [
          [1, 3, 224, 224]
        ], threshold: 3, topPredictEntries: 5),
        baseModel: baseModel) as OnnxResponsModel;

//restart model
await aiModelLand.restartModel(
        request: OnnxRequestModel(
        loadModelWay: LoadModelWay.fromAssets,
        onnxBackend: ONNXBackend.cpu),
        baseModel: baseModel);

//stop model
await aiModelLand.stopModel(baseModel: baseModel);


//delete model from device
await aiModelLand.deleteModelFromDevice(baseModel: baseModel);

//you can check information about device
await aiModelLand.checkPlatformInfo();

Running Examples

Instructions for running the included examples:

  1. Navigate to the example directory:
    cd example
  2. Run flutter run to launch the example app on your preferred device.
  3. Explore the example app to see how to use the library in your own project.

Contributing

Instructions for contributing to this library:

  1. Fork this repository.
  2. Create a new branch:
    git checkout -b my-new-branch
  3. Make changes and commit them:
    git commit -m "My message"
  4. Push to the remote branch:
    git push origin my-new-branch
  5. Create a pull request.

List any dependencies or prerequisites needed before getting started with editing this library:

  • Flutter
  • Installed Android Studio and Android SDK (For Android dev)
  • Installed Xcode (For iOS, macOS dev)

Instructions for building the library from source code:

  1. Clone this repository:
    git clone https://github.com/vlmoon99/ai_model_land
  2. Navigate to the project root directory:
    cd ai_model_land

About

No description, website, or topics provided.

Resources

Readme

License

View license
Activity

Stars

0 stars

Watchers

1 watching

Forks

1 fork
Report repository

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •  

Languages