Tensorflow t GCMLE上的预估器“；内部错误：作业“；大师；未在群集中定义；

我正试图使用GCMLE和TF版本1.8，并遵循相关说明

对于TF 1.8，他们说：

tpu_cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu,
zone=FLAGS.tpu_zone,
project=FLAGS.gcp_project)

config = tpu_config.RunConfig(
    cluster=tpu_cluster_resolver,
    model_dir=FLAGS.model_dir,
    save_checkpoints_steps=max(600, FLAGS.iterations_per_loop),
    tpu_config=tpu_config.TPUConfig(
        iterations_per_loop=FLAGS.iterations_per_loop,
        num_shards=FLAGS.num_cores))

然后我将其传递给TPUEstimator/train_and_evaluate（），如下所示：

estimator = tpu_estimator.TPUEstimator(
    use_tpu=True,
    model_fn=model_fn,
    config=run_config,
    params = params,
    train_batch_size = params.train_batch_size,
    eval_batch_size = params.eval_batch_size,
    )

tf.estimator.train_and_evaluate(estimator, train_spec, eval_spec)

对于1.7，他们分别表示使用带有“主”标志的不同配置。但是，当我在GCMLE上运行上述1.8指令，并将

--runtime_version

设置为1.8时，我得到以下回溯错误，这表明

内部错误：集群中未定义作业“master”

Traceback (most recent call last): File "/usr/lib/python2.7/runpy.py", line 174, in _run_module_as_main "__main__", fname, loader, pkg_name) File "/usr/lib/python2.7/runpy.py", line 72, in _run_code exec code in run_globals File "/root/.local/lib/python2.7/site-packages/trainer/task.py", line 585, in <module> run_experiment(params) File "/root/.local/lib/python2.7/site-packages/trainer/task.py", line 127, in run_experiment tf.estimator.train_and_evaluate(estimator, train_spec, eval_spec) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py", line 439, in train_and_evaluate executor.run() File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py", line 546, in run getattr(self, task_to_run)() File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py", line 601, in run_master self._start_distributed_training(saving_listeners=saving_listeners) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py", line 719, in _start_distributed_training self._start_std_server(config) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py", line 707, in _start_std_server start=False) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/server_lib.py", line 147, in __init__ self._server_def.SerializeToString(), status) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/errors_impl.py", line 519, in __exit__ c_api.TF_GetCode(self.status.status)) InternalError: Job "master" was not defined in cluster

Traceback（最近一次调用最后一次）：文件“/usr/lib/python2.7/runpy.py”，第174行，在运行模块中作为主“/uuuuu main”，fname，loader，pkg_name）文件“/usr/lib/python2.7/runpy.py”，第72行，在运行全局文件“/root/.local/lib/python2.7/site packages/task.py”，第585行，在运行实验（params）文件中“/root/.local/lib/python2.7/site packages/trainer/task.py”，第127行，在train\u实验tf.estimator.train\u和\u evaluate（estimator，train\u spec，eval\u spec）文件中“/usr/local/lib/python2.7/dist packages/tensorflow/python/estimator/training.py”，第439行，在train\u和\u evaluate executor.run（）文件中”/usr/local/lib/python2.7/dist packages/tensorflow/python/estimator/training.py”，第546行，运行getattr（self，task_to_run）（）文件“/usr/local/lib/python2.7/dist packages/tensorflow/python/estimator/training.py”，第601行，运行_master self.\u启动_分布式_培训（保存侦听器=保存侦听器）文件”/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py”，第719行，在“开始”分布式“培训自我”文件中，在“开始”std\u-server（配置）文件中“/usr/local/lib/python2.7/dist-packages/tensorflow/python/estimator/training.py”，第707行，在“开始”std\u-server-start=False）文件中”/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/server_-lib.py”，第147行，在uuu-init_uuu-self.u-server_-def.SerializeToString（），status）文件“/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/errors_impl.py”，第519行，在u-exit_u-api.TF-GetCode（self.status）内部错误：作业“主“未在群集中定义

---

这令人困惑，因为文档中说不要使用

master

，所以我不确定出了什么问题？

TPUEstimator是使用estimator.train（input\fn=train\u input\u fn，max\u steps=next\u checkpoint）和estimator.evaluate（）调用的

---

tf.estimator.train_和evaluate（estimator，train_spec，eval_spec）仅在CPU/GPU上工作。

指南已更新[1]。现在只有Cloud ML引擎运行时版本1.8到1.9可用。这些更新版本和最新的说明是否仍然存在问题？[1]