Problems using Google YAMNet for environmental sound classification in flutter

I discovered an error when testing on a real device. My module is based on modifying the sample from

https://www.tensorflow.org/tutorials/audio/transfer_learning_audio

and then converted to tflite. I followed the settings like xml, build.grade, yaml, and included labels.txt and tflite. But when I tested it on a real device, it could not run.

Parameters:

Syncing files to device SM A7050...
I/flutter (21612): Initializing state...
I/flutter (21612): Loading model...
D/TfliteAudio(21612): model name is: assets/converted_model.tflite
I/flutter (21612): Building widget...
I/InterpreterApi(21612): Loaded native library: tensorflowlite_jni
I/InterpreterApi(21612): Didn't load native library: tensorflowlite_jni_gms_client
I/tflite (21612): Initialized TensorFlow Lite runtime.
I/tflite (21612): Created TensorFlow Lite XNNPACK delegate for CPU.
D/TfliteAudio(21612): inputShape: [1]
D/TfliteAudio(21612): label name is: assets/labels.txt
D/TfliteAudio(21612): Labels: [crickets, crying_baby]
D/TfliteAudio(21612): loadModel parameters: {isAsset=true, outputRawScores=false, model=assets/converted_model.tflite, inputType=decodedWav, label=assets/labels.txt, numThreads=1}
I/SurfaceView(21612): applySurfaceTransforms: t = android.view.SurfaceControl$Transaction@3ede20b surfaceControl = Surface(name=SurfaceView - com.example.baby_cryr/com.example.baby_cryr.MainActivity@991d082@0)/@0x6aac8e8 frame = 1
I/ample.baby_cry(21612): WaitForGcToComplete blocked RunEmptyCheckpoint on ProfileSaver for 9.483ms
I/ViewRootImpl@d891c9cMainActivity: MSG_WINDOW_FOCUS_CHANGED 1 1
D/InputMethodManager(21612): prepareNavigationBarInfo() DecorView@9c71922[MainActivity]
D/InputMethodManager(21612): getNavigationBarColor() -16711423
D/InputMethodManager(21612): prepareNavigationBarInfo() DecorView@9c71922[MainActivity]
D/InputMethodManager(21612): getNavigationBarColor() -16711423
V/InputMethodManager(21612): Starting input: tba=com.example.baby_cryr ic=null mNaviBarColor -16711423 mIsGetNaviBarColorSuccess true , NavVisible : true , NavTrans : false
D/InputMethodManager(21612): startInputInner - Id : 0
I/InputMethodManager(21612): startInputInner - mService.startInputOrWindowGainedFocus
D/InputTransport(21612): Input channel constructed: 'ClientS', fd=107
D/InputMethodManager(21612): prepareNavigationBarInfo() DecorView@9c71922[MainActivity]
D/InputMethodManager(21612): getNavigationBarColor() -16711423
V/InputMethodManager(21612): Starting input: tba=com.example.baby_cryr ic=null mNaviBarColor -16711423 mIsGetNaviBarColorSuccess true , NavVisible : true , NavTrans : false
D/InputMethodManager(21612): startInputInner - Id : 0
I/SurfaceControl(21612): nativeRelease nativeObject s[-5476376610795376448]
I/SurfaceControl(21612): nativeRelease nativeObject e[-5476376610795376448]
I/SurfaceControl(21612): nativeRelease nativeObject s[-5476376610795376272]
I/SurfaceControl(21612): nativeRelease nativeObject e[-5476376610795376272]
I/ViewRootImpl@d891c9cMainActivity: ViewPostIme pointer 0
I/ViewRootImpl@d891c9cMainActivity: ViewPostIme pointer 1
I/flutter (21612): Requesting microphone permission...
I/flutter (21612): Microphone permission granted.
I/flutter (21612): Starting audio recognition...
D/TfliteAudio(21612): Parameters: {detectionThreshold=0.3, minimumTimeBetweenSamples=0, method=setAudioRecognitionStream, numOfInferences=1, averageWindowDuration=0, audioLength=0, sampleRate=16000, suppressionTime=0, bufferSize=2000}
D/TfliteAudio(21612): AudioLength has been readjusted. Length: 1
E/EventChannel#AudioRecognitionStream(21612): Failed to open event stream
E/EventChannel#AudioRecognitionStream(21612): java.lang.ArrayIndexOutOfBoundsException: length=1; index=1
E/EventChannel#AudioRecognitionStream(21612): at flutter.tflite_audio.TfliteAudioPlugin.determineAudio(TfliteAudioPlugin.java:296)
E/EventChannel#AudioRecognitionStream(21612): at flutter.tflite_audio.TfliteAudioPlugin.onListen(TfliteAudioPlugin.java:244)
E/EventChannel#AudioRecognitionStream(21612): at io.flutter.plugin.common.EventChannel$IncomingStreamRequestHandler.onListen(EventChannel.java:218)
E/EventChannel#AudioRecognitionStream(21612): at io.flutter.plugin.common.EventChannel$IncomingStreamRequestHandler.onMessage(EventChannel.java:197)
E/EventChannel#AudioRecognitionStream(21612): at io.flutter.embedding.engine.dart.DartMessenger.invokeHandler(DartMessenger.java:295)
E/EventChannel#AudioRecognitionStream(21612): at io.flutter.embedding.engine.dart.DartMessenger.lambda$dispatchMessageToQueue$0$io-flutter-embedding-engine-dart-DartMessenger(DartMessenger.java:322)
E/EventChannel#AudioRecognitionStream(21612): at io.flutter.embedding.engine.dart.DartMessenger$$ExternalSyntheticLambda0.run(Unknown Source:12)
E/EventChannel#AudioRecognitionStream(21612): at android.os.Handler.handleCallback(Handler.java:938)
E/EventChannel#AudioRecognitionStream(21612): at android.os.Handler.dispatchMessage(Handler.java:99)
E/EventChannel#AudioRecognitionStream(21612): at android.os.Looper.loop(Looper.java:246)
E/EventChannel#AudioRecognitionStream(21612): at android.app.ActivityThread.main(ActivityThread.java:8653)
E/EventChannel#AudioRecognitionStream(21612): at java.lang.reflect.Method.invoke(Native Method)
E/EventChannel#AudioRecognitionStream(21612): at com.android.internal.os.RuntimeInit$MethodAndArgsCaller.run(RuntimeInit.java:602)
E/EventChannel#AudioRecognitionStream(21612): at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:1130)

======== Exception caught by services library ======================================================
The following PlatformException was thrown while activating platform stream on channel AudioRecognitionStream:
PlatformException(error, length=1; index=1, null, null)

When the exception was thrown, this was the stack:

#0 StandardMethodCodec.decodeEnvelope (package:flutter/src/services/message_codecs.dart:652:7)
#1 MethodChannel._invokeMethod (package:flutter/src/services/platform_channel.dart:310:18)
#2 EventChannel.receiveBroadcastStream.
(package:flutter/src/services/platform_channel.dart:652:9) ====================================================================================================

I was hoping you might be able to help me understand whether the issue is with my trained model or if the platform simply does not support it. It would be greatly appreciated if you could provide some guidance. Thank you for your time and assistance

[Caldarie]

hi @tommygor20,

It seems you have set the input shape to decodedWav, which is incompatible with Yamnet.

I have been meaning to make this plugin compatible with yamnet, but just havent had much time to do so. If you are keen on the idea, would you be open to adding that feature to the plugin?