Tensorflow Lite model in wrong output format how to use in Google ML Kit?

Question

I want to use cameraX to detect live stream to accomplish object detection, so I use ML Kit .

I'm failed to put image into my model.

my model's

inputs:

• Type : Image<float32>

• shape: [1,640,640,3]

• Min/Max :[-1] / [1]

outputs:

• Type : BoundingBox<float32>

• shape: [1,9,8400]

• Min/Max :[] / []

mlkit expected output

below is my error message and related code:

Pass image to an ML Kit failed:                                                                                                com.google.mlkit.common.MlKitException: Failed to initialize detector. Unexpected number of dimensions for output index 0: got 3D, expected either 2D (BxN with B=1) or 4D (BxHxWxN with B1, W=1, H=1).

val customObjectDetectorOptions =
    CustomObjectDetectorOptions.Builder(localModel)
        .setDetectorMode(CustomObjectDetectorOptions.STREAM_MODE)
        .enableClassification()
        .setClassificationConfidenceThreshold(0.5f)
        .setMaxPerObjectLabelCount(3)
        .build()

val objectDetector =
    ObjectDetection.getClient(customObjectDetectorOptions)

 private class ImageAnalyzer : ImageAnalysis.Analyzer {
        @OptIn(ExperimentalGetImage::class) override fun analyze(imageProxy: ImageProxy) {
        //@OptIn(ExperimentalGetImage::class) override fun analyze(imageProxy: ImageProxy) {

            val mediaImage = imageProxy.image
            if (mediaImage != null) {
                Log.d(TAG, "mediaImage != null: succeed")
               
                val image = InputImage.fromMediaImage(mediaImage, imageProxy.imageInfo.rotationDegrees)

         
                // Pass image to an ML Kit Vision API
                objectDetector.process(image)
                    .addOnSuccessListener { ...}

I guess I need to transform my model's output shape from (1,9,8400) to 2D or 4D dimension,but I have no idea how to do . Hope someone can help me!!!