Error while deploying ML model in Android Kotlin

TensorFlow General Discussion
,
1

My model takes images and a np.ndarray

Below is the Sample code generated by Android Studio in Kotlin language after importing my model.

    val model = ConvertedModel.newInstance(context)
    
    // Creates inputs for reference.
    val inputFeature0 = TensorBuffer.createFixedSize(intArrayOf(1, 1), DataType.FLOAT32)
    inputFeature0.loadBuffer(byteBuffer)
    val inputFeature1 = TensorBuffer.createFixedSize(intArrayOf(1, 1, 1, 3), DataType.FLOAT32)
    inputFeature1.loadBuffer(byteBuffer)
    ...
    ...
    // Runs model inference and gets result.
    val outputs = model.process(inputFeature0, inputFeature1, inputFeature2)
    ...
    ...
    // Releases model resources if no longer used.
    model.close()

I have taken 2 images and converted into bitmap

    var tensorImage0 = TensorImage(DataType.FLOAT32)
                    tensorImage0.load(my_img_bitmap)
    ...    
    val batchBuff = TensorBuffer.createFixedSize(intArrayOf(1, 1), DataType.FLOAT32)

Added additional processing

    var imageProcessor = ImageProcessor.Builder()
            .add(ResizeOp(1, 1, ResizeOp.ResizeMethod.BILINEAR))
            .build()

    tensorImage0 = imageProcessor.process(tensorImage0)
    ...

Then passed it into the input buffer

    val inputFeature0 = TensorBuffer.createFixedSize(intArrayOf(1, 1), DataType.FLOAT32)
    inputFeature0.loadBuffer(batchBuff.buffer)
    val inputFeature1 = TensorBuffer.createFixedSize(intArrayOf(1, 1, 1, 3), DataType.FLOAT32)
    inputFeature1.loadBuffer(tensorImage0.buffer)
    ...

When this process starts the app crashes.

#00 pc 000000000005e03c /system/lib64/libbinder.so (android::IPCThreadState::transact(int, unsigned int, android::Parcel const&, android::Parcel*, unsigned int)+248)

Please help me overcome the issue.