我正在尝试运行一个自定义的Tflie模型在Android的分割。它在CPU上运行良好，但在使用GPU委托时，它崩溃并出现以下错误：

E/AndroidRuntime(30867): FATAL EXCEPTION: AsyncTask #2
E/AndroidRuntime(30867): Process: com.apptec.scansdk, PID: 30867
E/AndroidRuntime(30867): java.lang.RuntimeException: An error occurred while executing doInBackground()
E/AndroidRuntime(30867):    at android.os.AsyncTask$3.done(AsyncTask.java:309)
E/AndroidRuntime(30867):    at java.util.concurrent.FutureTask.finishCompletion(FutureTask.java:354)
E/AndroidRuntime(30867):    at java.util.concurrent.FutureTask.setException(FutureTask.java:223)
E/AndroidRuntime(30867):    at java.util.concurrent.FutureTask.run(FutureTask.java:242)
E/AndroidRuntime(30867):    at android.os.AsyncTask$SerialExecutor$1.run(AsyncTask.java:234)
E/AndroidRuntime(30867):    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1113)
E/AndroidRuntime(30867):    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:588)
E/AndroidRuntime(30867):    at java.lang.Thread.run(Thread.java:818)
E/AndroidRuntime(30867): Caused by: java.lang.IllegalArgumentException: Internal error: Failed to run on the given Interpreter: TensorDescriptor don't have selector with name - Slice
E/AndroidRuntime(30867): Falling back to OpenGL
E/AndroidRuntime(30867): TfLiteGpuDelegate Invoke: GpuDelegate must run on the same thread where it was initialized.
E/AndroidRuntime(30867): Node number 105 (TfLiteGpuDelegateV2) failed to invoke.
E/AndroidRuntime(30867): 
E/AndroidRuntime(30867):    at org.tensorflow.lite.NativeInterpreterWrapper.run(Native Method)
E/AndroidRuntime(30867):    at org.tensorflow.lite.NativeInterpreterWrapper.run(NativeInterpreterWrapper.java:158)
E/AndroidRuntime(30867):    at org.tensorflow.lite.Interpreter.runForMultipleInputsOutputs(Interpreter.java:343)
E/AndroidRuntime(30867):    at org.tensorflow.lite.Interpreter.run(Interpreter.java:304)
E/AndroidRuntime(30867):    at com.apptec.scansdk.MainActivity$RunSegmentationOnBinary.runTflite(MainActivity.java:237)
E/AndroidRuntime(30867):    at com.apptec.scansdk.MainActivity$TfliteTask.doInBackground(MainActivity.java:310)
E/AndroidRuntime(30867):    at com.apptec.scansdk.MainActivity$TfliteTask.doInBackground(MainActivity.java:284)
E/AndroidRuntime(30867):    at android.os.AsyncTask$2.call(AsyncTask.java:295)
E/AndroidRuntime(30867):    at java.util.concurrent.FutureTask.run(FutureTask.java:237)
E/AndroidRuntime(30867):    ... 4 more

下面是我的代码加载模型

MappedByteBuffer buffer = null;
        buffer = FileUtil.loadMappedFile(context, "docScan.tflite");
int numThreads = 4;
boolean useGpuDelegate = true;
final Interpreter.Options tfliteOptions = new Interpreter.Options();
tfliteOptions.setNumThreads(numThreads);
if (useGpuDelegate){
    GpuDelegate delegate = new GpuDelegate();
    tfliteOptions.addDelegate(delegate);
}
tfLite = new Interpreter(buffer, tfliteOptions);

奇怪的是，它适用于MI/Redmi/OPPO等手机，只在三星上崩溃。救命啊！