python之分布式 tensorflow : Create a session in only one worker to print the results

我只需要打印在训练中分配的一些全局变量的值。关闭“MonitoredTrainingSession”后，我仅使用以下方法在 chief worker 中创建了一个 session :

    if FLAGS.task_index == 0: 
        with tf.Session() as sess: 
            print sess.run(some_variable) 

但是我一直收到错误:

InvalidArgumentError (see above for traceback): Cannot assign a device to node 'save/RestoreV2_20': Could not satisfy explicit device specification '/job:ps/task:0/device:CPU:0' because no devices matching that specification are registered in this process; available devices: /job:localhost/replica:0/task:0/cpu:0

知道怎么做吗？

请您参考如下方法:

available devices: /job:localhost/replica:0/task:0/cpu:0

建议您使用您创建的服务器。尝试在创建 session 时传递 server.target。

with tf.Session(server.target) as sess: 
    print sess.run(some_variable)