Tensorflow lite with own model

[vbpo...@gmail.com]

I am learning about object detection of tensorflowlite. When I put my own training model into the model with parameters such as:

+ inputSize: 128 (128x128) is different from sample is 300x300

+ NUM_DETECTIONS: 2 (detects up to 2 objects) other than the sample of 10

+ and training photos are gray images

When I ran the project, the following error occurred. I have found out how to fix bugs online but not feasible. Anyone who has knowledge about this area can help me find a problem.

2019-11-28 09:11:14.488 29195-29339/com.example.driscanai E/AndroidRuntime: FATAL EXCEPTION: inference

    Process: com.example.driscanai, PID: 29195

    java.lang.IllegalArgumentException: Cannot copy between a TensorFlowLite tensor with shape [1, 64, 3, 7] and a Java object with shape [1, 2, 4].

        at org.tensorflow.lite.Tensor.throwIfShapeIsIncompatible(Tensor.java:342)

        at org.tensorflow.lite.Tensor.throwIfDataIsIncompatible(Tensor.java:305)

        at org.tensorflow.lite.Tensor.copyTo(Tensor.java:183)

        at org.tensorflow.lite.NativeInterpreterWrapper.run(NativeInterpreterWrapper.java:164)

        at org.tensorflow.lite.Interpreter.runForMultipleInputsOutputs(Interpreter.java:296)

        at com.example.driscanai.utils.custom.TFLiteObjectDetectionAPIModel.recognizeImage(TFLiteObjectDetectionAPIModel.java:132)

        at com.example.driscanai.screen.camera.CameraActivity$5.run(CameraActivity.java:298)

        at android.os.Handler.handleCallback(Handler.java:790)

        at android.os.Handler.dispatchMessage(Handler.java:99)

        at android.os.Looper.loop(Looper.java:164)

        at android.os.HandlerThread.run(HandlerThread.java:65)