30

Tensorflow 错误集锦

 6 years ago
source link: http://whatbeg.com/2018/12/05/tensorflowtips.html?amp%3Butm_medium=referral
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
neoserver,ios ssh client 
2018-12-05 22:18:24.565303: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:222] Initialize GrpcChannelCache for job ps -> {0 -> localhost:3376}
2018-12-05 22:18:24.565372: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:222] Initialize GrpcChannelCache for job worker -> {0 -> localhost:3330, 1 -> localhost:3331}
2018-12-05 22:18:24.569212: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:381] Started server with target: grpc://localhost:3376
2018-12-05 22:18:26.170901: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:222] Initialize GrpcChannelCache for job ps -> {0 -> localhost:3376}
2018-12-05 22:18:26.170969: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:222] Initialize GrpcChannelCache for job worker -> {0 -> localhost:3330, 1 -> localhost:3331}
2018-12-05 22:18:26.174856: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:381] Started server with target: grpc://localhost:3330
2018-12-05 22:18:27.177003: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:222] Initialize GrpcChannelCache for job ps -> {0 -> localhost:3376}
2018-12-05 22:18:27.177071: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:222] Initialize GrpcChannelCache for job worker -> {0 -> localhost:3330, 1 -> localhost:3331}
2018-12-05 22:18:27.180980: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:381] Started server with target: grpc://localhost:3331
2018-12-05 22:18:34.625459: I tensorflow/core/distributed_runtime/master.cc:267] CreateSession still waiting for response from worker: /job:worker/replica:0/task:0
2018-12-05 22:18:34.625513: I tensorflow/core/distributed_runtime/master.cc:267] CreateSession still waiting for response from worker: /job:worker/replica:0/task:1
2018-12-05 22:18:36.231936: I tensorflow/core/distributed_runtime/master.cc:267] CreateSession still waiting for response from worker: /job:ps/replica:0/task:0
2018-12-05 22:18:36.231971: I tensorflow/core/distributed_runtime/master.cc:267] CreateSession still waiting for response from worker: /job:worker/replica:0/task:1
2018-12-05 22:18:37.235899: I tensorflow/core/distributed_runtime/master.cc:267] CreateSession still waiting for response from worker: /job:ps/replica:0/task:0
2018-12-05 22:18:37.235952: I tensorflow/core/distributed_runtime/master.cc:267] CreateSession still waiting for response from worker: /job:worker/replica:0/task:0

首先保证 job_name,task_index,ps_hosts,worker_hosts 这四个参数都是正确的,考虑以下这种情况是不正确的：

在一个IP为192.168.1.100的机器上启动ps或worker进程：

--job_name=worker
--task_index=1
--ps_hosts=192.168.1.100:2222,192.168.1.101:2222
--worker_hosts=192.168.1.100:2223,192.168.1.101:2223

因为该进程启动位置是192.168.1.100，但是运行参数中指定的task_index为1，对应的IP地址是ps_hosts或worker_hosts的第二项（第一项的task_index为0)，也就是192.168.1.101，和进程本身所在机器的IP不一致。

另外一种情况也会导致该问题的发生，从TensorFlow-1.4开始，分布式会自动使用环境变量中的代理去连接，如果运行的节点之间不需要代理互连，那么将代理的环境变量移除即可，在脚本的开始位置添加代码：

注意这段代码必须写在import tensorflow as tf或者import moxing.tensorflow as mox之前

import os
os.enrivon.pop('http_proxy')
os.enrivon.pop('https_proxy')

— 摘自( https://bbs.huaweicloud.com/blogs/463145f7a1d111e89fc57ca23e93a89f )


report

Recommend

About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK