请问mode=='evaluate'时，run_evaluate函数里计算的metric只是一个batch的metric？

请问mode=='evaluate'时，run_evaluate函数里计算的metric只是一个batch的metric 吗？
如下代码，
with tf.train.MonitoredTrainingSession(
master=master, is_chief=is_chief,
checkpoint_dir=flags_obj.model_dir,
save_checkpoint_secs=None,
log_step_count_steps=None,
hooks=hooks, config=config) as sess:
while not sess.should_stop():
metric_val = sess.run(metric)

print('{}: {}'.format(metric_name, metric_val))
里面没有加hook，所以source没有next时自动停止吗？
每一次batch，返回的metric都是一个具体值，这样最后metric_val 就是最后一个batch计算出的
metric，请问是这样的吗？

Read the run_evaluate implementation and the MonitoredTrainingSession loop shown in the issue, then trace how metric and source are defined. Run the evaluate path and inspect whether metric_val represents one batch or an aggregate, and whether iteration stops when source has no next element; done means the behavior is documented or corrected with a focused test.

Written by the indexing model from the issue text.