Write summaries to be viewed by TensorBoard. Args: job_name: The name of this job ('trainer', 'evaler', etc.) dataset_name: The dataset name that's decoded. global_step: Integer number of trainer steps (not a tensor). summaries: Dict of {summary_name: tf.Summary()}.
(self, job_name, dataset_name, global_step, summaries)
source not stored for this graph (policy: none)
no test coverage detected