Python google ai平台上的超参数调整错误：“0”；副本主机0以非零状态1“退出；

在google cloud AI平台中使用hyperparameter tuning（我的hyperparameter配置信息在YAML文件中）培训深度学习模型时，我遇到以下错误：

Hyperparameter Tuning Trial #2 Failed before any other successful trials were completed. 
The failed trial had parameters: batch_size=11, learning_rate=3.527059074944887e-05, .
The trial's error message was: The replica master 0 exited with a non-zero status of 1

由于错误消息有点泛化，我很难理解问题出在哪里

我的YAML配置文件：

trainingInput:
  hyperparameters:
    goal: MINIMIZE
    maxTrials: 2
    maxParallelTrials: 2
    hyperparameterMetricTag: loss
    enableTrialEarlyStopping: FALSE
    params:
      - parameterName: batch_size
        type: INTEGER
        minValue: 8
        maxValue: 16
        scaleType: UNIT_LINEAR_SCALE
      - parameterName: learning_rate
        type: DOUBLE
        minValue: 0.00001
        maxValue: 0.0001
        scaleType: UNIT_LINEAR_SCALE

---

您能否提供有关错误消息的更详细日志？检查错误的回溯日志。检查回溯日志（）