在TensorFlow Lite中没有从机器模型获得任何输出

如何解决在TensorFlow Lite中没有从机器模型获得任何输出

我正在使用TensorFlow的Keras API开发机器模型，该模型可以根据输入数据准确预测正确的字母。训练和测试数据是分压器中3个挠性传感器的电压值；接收到的电压是浮子（X.XX），用于训练顺序模型。我已经将文件从python转换为.h文件，并将其放置在Arduino草图所在的文件夹中。

https://www.tensorflow.org/lite/microcontrollers

我目前正在尝试将其集成到Arduino Nano 33 BLE Sense中。我希望从输出串行监视器中看到一个可能的预测类，然后是其旁边的百分比或概率。可能的预测类是“ A”，“ B”，“ C”，“ D”和“ RELAXED”。但是，我没有从串行监视器获得任何输出。我认为问题出在我将输入数据发送到输入缓冲区的方式上，或者可能是因为其格式错误。请，如果您可以看一下我的代码，我将不胜感激。谢谢。

我的代码：

/* 
 *  Flex Classifier which is to test how the model is integrated into arduino
 *  
 */
////////////////////MODEL////////////////////
// Including Libraries for TensorFlow
#include <TensorFlowLite.h>
#include <tensorflow/lite/micro/all_ops_resolver.h>
#include <tensorflow/lite/micro/micro_error_reporter.h>
#include <tensorflow/lite/micro/micro_interpreter.h>
#include <tensorflow/lite/schema/schema_generated.h>
#include <tensorflow/lite/version.h>

// Including Model
#include "Letter_Model5.h"

// global variables used for TensorFlow Lite (Micro)
tflite::MicroErrorReporter tflErrorReporter;

// pull in all the TFLM ops,you can remove this line and
// only pull in the TFLM ops you need,if would like to reduce
// the compiled size of the sketch.
tflite::AllOpsResolver tflOpsResolver;

const tflite::Model* tflModel = nullptr;
tflite::MicroInterpreter* tflInterpreter = nullptr;
TfLiteTensor* tflInputTensor = nullptr;
TfLiteTensor* tflOutputTensor = nullptr;

// Create a static memory buffer for TFLM,the size may need to
// be adjusted based on the model you are using
constexpr int tensorArenaSize = 8 * 1024;
byte tensorArena[tensorArenaSize];

// array to map gesture index to a name
 const char* GESTURES[] = {
  "A","B","C","D","RELAXED"
};

#define NUM_GESTURES (sizeof(GESTURES) / sizeof(GESTURES[0]))

////////////////////BOARD////////////////////

const float V = 3.30; //Input voltage on board

////////////////////FLEX////////////////////

//Flex Sensor #1
const int FLEX_PIN1 = A0;
//Flex Sensor #2
const int FLEX_PIN2 = A1;
//Flex Sensor #3
const int FLEX_PIN3 = A2;

void setup() 
{
  // get the TFL representation of the model byte array
  tflModel = tflite::GetModel(Letter_Model5);
  if (tflModel->version() != TFLITE_SCHEMA_VERSION) 
  {
    Serial.println("Model schema mismatch!");
    while (1);
  }

  // Create an interpreter to run the model
  tflInterpreter = new tflite::MicroInterpreter(tflModel,tflOpsResolver,tensorArena,tensorArenaSize,&tflErrorReporter);

  // Allocate memory for the model's input and output tensors
  tflInterpreter->AllocateTensors();

  // Get pointers for the model's input and output tensors
  tflInputTensor = tflInterpreter->input(0);
  tflOutputTensor = tflInterpreter->output(0);

  // Begin Serial
  Serial.begin(9600);
  while (!Serial); // Wait for the Serial to start

  // Defining flex ports
  pinMode(FLEX_PIN1,INPUT);
  pinMode(FLEX_PIN2,INPUT);
  pinMode(FLEX_PIN3,INPUT);

}

void loop() {
  // Reading flex inputs
  int flexADC1 = analogRead(FLEX_PIN1);
  int flexADC2 = analogRead(FLEX_PIN2);
  int flexADC3 = analogRead(FLEX_PIN3);

  // Finding Voltages from flex inputs
  float flexV1  = (flexADC1/1023.0)*V;
  float flexV2  = (flexADC2/1023.0)*V;
  float flexV3  = (flexADC3/1023.0)*V;

  //Sending input data into the model
  float InputData[] = {flexV1,flexV2,flexV3};
  tflInputTensor->data.f[0] = InputData[0];
  tflInputTensor->data.f[1] = InputData[1];
  tflInputTensor->data.f[2] = InputData[2];

  //Printing out predicted string from the output tensor
  for (int i = 0; i < NUM_GESTURES; i++)
  {
    Serial.print(GESTURES[i]);
    Serial.print(": ");
    Serial.println(tflOutputTensor->data.f[i],2);
  }
  delay(500);
}

这就是我在串行监视器上获得的输出。另外，为什么字母“ B”得到“ nan”？ https://imgur.com/a/eaSBLBe