Flutter plugin for speech recognition and transcription using DeepSpeech.
Step 1:
Add the plugin as a dependency in your pubspec.yaml
flutter_deep_speech:
git:
url: git://github.com/AdarshaBista/FlutterDeepSpeech.git
Step 2:
Add permission for microphone in AndroidManifest.xml
<uses-permission android:name="android.permission.RECORD_AUDIO" />
Step 3: Use the plugin. See the example app for the complete usage.
Note:
You need to obtain a tflite
model and scorer
file to use the plugin. You can download the files from the DeepSpeech repository. Once downloaded, place the files inside assets/models/
directory of your project. The models are not loaded from the assets directory directly, but are copied to the app's external files directory first and then loaded from there (See the example app). The copying of the files is not handled by the plugin.
See the Changelog for supported features and limitations.
Android version referenced from android_mic_streaming example.