TensorBoard 使用和问题解决

TensorBoard 使用和问题解决

一、启动TensorBoard　

1)

python tensorflow/tensorboard/tensorboard.py --logdir=path/to/log-directory

logdir 指向 SummaryWriter 序列化数据的存储路径。
通过在浏览器中输入 localhost:6006 来查看 TensorBoard。

2)pip安装了 TensorBoard

tensorboard --logdir=/path/to/log-directory

二、TensorBoard 问题　

1）问题：ImportError: cannot import name run_main

dyufei@nodeB:~/source/test/note$ sudo tensorboard --logdir="/home/dyufei/source/test/note/my_graph"
Traceback (most recent call last):
File "/usr/local/bin/tensorboard", line 7, in
from tensorboard.main import run_main
ImportError: cannot import name run_main

解决重新pip安装 tensorboard

sudo pip install tensorboard

2）问题：启动在服务器上后无法远程查看

Starting TensorBoard 41 on port 6006
(You can navigate to http://127.0.1.1:6006)

解决：启动时加--host 指定ip

tensorboard --host=10.10.101.2 --logdir="my_graph"

3）问题：端口被占用

ERROR:tensorflow:Tried to connect to port 6006, but address is in use.
Tried to connect to port 6006, but address is in use.

解决：--port 指定端口

tensorboard --host=10.10.101.2 --port=6099 --logdir="my_graph"

转载于:https://www.cnblogs.com/dyufei/p/8075006.html