Android摄像头帧年龄性别识别中GpuDelegate初始化报错求助

错误信息 #

E/skurt.genderag: [qarth_debug:]  get PatchStore::createDisableExceptionQarthFile method fail.
E/AndroidRuntime: FATAL EXCEPTION: main
    Process: com.esrabaskurt.genderage, PID: 11140
    java.lang.NoClassDefFoundError: Failed resolution of: Lorg/tensorflow/lite/gpu/GpuDelegateFactory$Options;
        at org.tensorflow.lite.gpu.GpuDelegate.<init>(GpuDelegate.java:52)
        at com.esrabaskurt.genderage.age_gender_recognition.<init>(age_gender_recognition.java:56)
        at com.esrabaskurt.genderage.CameraActivity.onCreate(CameraActivity.java:79)
        ...（省略后续栈信息）
I/Process: Sending signal. PID: 11140 SIG: 9

相关代码片段 #

age_gender_recognition.java 第56行 #

import org.tensorflow.lite.gpu.GpuDelegate;
public class age_gender_recognition {
...
    private GpuDelegate gpuDelegate=null;
    ...
    age_gender_recognition(AssetManager assetManager, Context context,String modelPath,int inputSize) throws IOException
    {
...
    gpuDelegate =new GpuDelegate();
...
    }}

CameraActivity.java 第79行 #

age_gender_recognition =new age_gender_recognition(getAssets(),CameraActivity.this,"model.tflite",inputSize);

build.gradle 依赖配置 #

dependencies {
    implementation 'org.tensorflow:tensorflow-lite-metadata:0.4.3'
    implementation 'org.tensorflow:tensorflow-lite-gpu:2.11.0'
    implementation 'org.tensorflow:tensorflow-lite-support:0.4.3'
    implementation 'org.tensorflow:tensorflow-lite-task-vision:0.4.3'

    implementation 'org.tensorflow:tensorflow-lite-task-text:0.4.3'
...
}

解决建议 #

• 统一TensorFlow Lite依赖版本
  当前依赖中tensorflow-lite-gpu版本为2.11.0，其他TFLite相关依赖版本为0.4.3，版本不匹配会导致类加载失败。确保所有TensorFlow Lite相关依赖版本对齐，例如保持0.4.3（该版本对应TFLite 2.11.0），若升级则同步所有相关依赖版本。

• 修正GpuDelegate初始化方式
  TFLite 2.11.0中GpuDelegate无参构造内部依赖已变更，建议使用带配置项的构造方法：

  import org.tensorflow.lite.gpu.GpuDelegate;
  
  public class age_gender_recognition {
      private GpuDelegate gpuDelegate = null;
  
      age_gender_recognition(AssetManager assetManager, Context context, String modelPath, int inputSize) throws IOException {
          GpuDelegate.Options options = new GpuDelegate.Options();
          // 根据需求配置参数，示例配置如下
          options.setPrecisionLossAllowed(true);
          options.setInferencePreference(GpuDelegate.Options.INFERENCE_PREFERENCE_FAST_SINGLE_ANSWER);
          gpuDelegate = new GpuDelegate(options);
          // 其他初始化代码
      }
  }
  

• 清理重建项目
  执行Build > Clean Project，再执行Build > Rebuild Project，清除应用缓存后重新安装。

• 检查多Dex配置
  若项目启用多Dex，确保GpuDelegateFactory$Options被包含在主Dex中：

  android {
      defaultConfig {
          multiDexEnabled true
      }
  }
  dependencies {
      implementation 'androidx.multidex:multidex:2.0.1'
  }
  

内容的提问来源于stack exchange，提问作者esra