创建MonitoredTrainingSession会导致InvalidArgumeInteror以十为单位

Question

用于启动MonitoredTrainingSession的代码导致InvalidArgumentError：

with tf.train.MonitoredTrainingSession(
    server.target,
    is_chief=is_learner,
    checkpoint_dir=FLAGS.logdir,
    save_checkpoint_secs=600,
    save_summaries_secs=30,
    log_step_count_steps=50000,
    hooks=[py_process.PyProcessHook()],
    config=config) as session:

以下是回溯：

Traceback (most recent call last):
  File "/Users/hughalessi/miniconda3/envs/rl/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1334, in _do_call
    return fn(*args)
  File "/Users/hughalessi/miniconda3/envs/rl/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1317, in _run_fn
    self._extend_graph()
  File "/Users/hughalessi/miniconda3/envs/rl/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1352, in _extend_graph
    tf_session.ExtendSession(self._session)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Input 1 of node scan/while/Merge_5 was passed double from scan/while/NextIteration_5:0 incompatible with expected float.

这是一个使用定制环境构建的可伸缩的\u代理，谷歌的IMPALA架构的实现。以下是指向存储库的链接： https://github.com/deepmind/scalable_agent