Skip to content

Tensorboard does not show any scalers or graphs  #9701

Closed
@MMCMA

Description

@MMCMA

System information

Have I written custom code (as opposed to using a stock example script provided in TensorFlow): yes
OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Win10
TensorFlow installed from (source or binary): binary, installed via pip
TensorFlow version (use command below): 1.1.0
Bazel version (if compiling from source):na
CUDA/cuDNN version: na
GPU model and memory: na
Exact command to reproduce:?

Describe the problem

Initially I was working with tensorflow 1.0.0 and tensorboard seemed to work just fine. I have updated to 1.1.0 and now it appears it does not display anything anymore. However, I get the warning 'WARNING:tensorflow:path ../external\data/plugin/text/runs not found, sending 404'. I have also downgraded tensorflow to 1.0.0 and now it also stopped working.

Source code / logs

tensorboard --inspect --logdir=model_dir
Found event files in: model_dir

These tags are in model_dir:
audio -
histograms
   dnn/dnn/hiddenlayer_0_activation
   dnn/dnn/hiddenlayer_10_activation
   dnn/dnn/hiddenlayer_11_activation
   dnn/dnn/hiddenlayer_1_activation
   dnn/dnn/hiddenlayer_2_activation
   dnn/dnn/hiddenlayer_3_activation
   dnn/dnn/hiddenlayer_4_activation
   dnn/dnn/hiddenlayer_5_activation
   dnn/dnn/hiddenlayer_6_activation
   dnn/dnn/hiddenlayer_7_activation
   dnn/dnn/hiddenlayer_8_activation
   dnn/dnn/hiddenlayer_9_activation
   dnn/dnn/logits_activation
images -
scalars
   dnn/dnn/hiddenlayer_0_fraction_of_zero_values
   dnn/dnn/hiddenlayer_10_fraction_of_zero_values
   dnn/dnn/hiddenlayer_11_fraction_of_zero_values
   dnn/dnn/hiddenlayer_1_fraction_of_zero_values
   dnn/dnn/hiddenlayer_2_fraction_of_zero_values
   dnn/dnn/hiddenlayer_3_fraction_of_zero_values
   dnn/dnn/hiddenlayer_4_fraction_of_zero_values
   dnn/dnn/hiddenlayer_5_fraction_of_zero_values
   dnn/dnn/hiddenlayer_6_fraction_of_zero_values
   dnn/dnn/hiddenlayer_7_fraction_of_zero_values
   dnn/dnn/hiddenlayer_8_fraction_of_zero_values
   dnn/dnn/hiddenlayer_9_fraction_of_zero_values
   dnn/dnn/logits_fraction_of_zero_values
   loss
tensor -
======================================================================

Event statistics for model_dir:
audio -
graph
   first_step           0
   last_step            0
   max_step             0
   min_step             0
   num_steps            1
   outoforder_steps     []
histograms
   first_step           1
   last_step            1
   max_step             1
   min_step             1
   num_steps            1
   outoforder_steps     []
images -
scalars
   first_step           1
   last_step            1
   max_step             1
   min_step             1
   num_steps            1
   outoforder_steps     []
sessionlog:checkpoint
   first_step           1
   last_step            1
   max_step             1
   min_step             1
   num_steps            1
   outoforder_steps     []
sessionlog:start
   outoforder_steps     []
   steps                [1]
sessionlog:stop -
tensor -
======================================================================


tensorboard --logdir=model_dir --host=127.0.0.1
Starting TensorBoard b'47' at http://127.0.0.1:6006
(Press CTRL+C to quit)
WARNING:tensorflow:path ../external\data/plugin/text/runs not found, sending 404
WARNING:tensorflow:path ../external\data/plugin/text/runs not found, sending 404
WARNING:tensorflow:path ../external\data/plugin/text/runs not found, sending 404
WARNING:tensorflow:path ../external\data/plugin/text/runs not found, sending 404

Activity

drpngx

drpngx commented on May 6, 2017

@drpngx
Contributor

Just checking that it's not a windows issue, @guschmue @vit-stepanovs do you guys run tensorboard on windows and is it working well for you?

guschmue

guschmue commented on May 7, 2017

@guschmue
Contributor

Works fine on master and 1.1 as far I can tell.
Above log shows tensorboard finds one 1 scalar, maybe the data in the scalar is not valid ?

MMCMA

MMCMA commented on May 7, 2017

@MMCMA
Author

I have tried the same for a different dir which used to work but no longer returns anything. here is the log of inspect:

C:\Users\T430MM>tensorboard --inspect --logdir=model_dir
======================================================================
Processing event files... (this can take a few minutes)
======================================================================

Found event files in:
model_dir
model_dir\eval

These tags are in model_dir:
audio -
histograms
   dnn/hiddenlayer_0_activation
   dnn/hiddenlayer_1_activation
   dnn/hiddenlayer_2_activation
   dnn/hiddenlayer_3_activation
   dnn/logits_activation
images -
scalars
   dnn/hiddenlayer_0_fraction_of_zero_values
   dnn/hiddenlayer_1_fraction_of_zero_values
   dnn/hiddenlayer_2_fraction_of_zero_values
   dnn/hiddenlayer_3_fraction_of_zero_values
   dnn/logits_fraction_of_zero_values
   global_step/sec
   loss
tensor -
======================================================================

Event statistics for model_dir:
audio -
graph
   first_step           0
   last_step            0
   max_step             0
   min_step             0
   num_steps            1
   outoforder_steps     []
histograms
   first_step           1
   last_step            9901
   max_step             9901
   min_step             1
   num_steps            100
   outoforder_steps     []
images -
scalars
   first_step           1
   last_step            9901
   max_step             9901
   min_step             1
   num_steps            100
   outoforder_steps     []
sessionlog:checkpoint
   first_step           1
   last_step            10000
   max_step             10000
   min_step             1
   num_steps            2
   outoforder_steps     []
sessionlog:start
   outoforder_steps     []
   steps                [1]
sessionlog:stop -
tensor -
======================================================================

These tags are in model_dir\eval:
audio -
histograms -
images -
scalars
   loss
tensor -
======================================================================

Event statistics for model_dir:
audio -
graph
   first_step           0
   last_step            0
   max_step             0
   min_step             0
   num_steps            1
   outoforder_steps     []
histograms -
images -
scalars
   first_step           10000
   last_step            10000
   max_step             10000
   min_step             10000
   num_steps            1
   outoforder_steps     []
sessionlog:checkpoint -
sessionlog:start -
sessionlog:stop -
tensor -
======================================================================

The browser loads the tensorboard page, but it displays the following error under the scalar tab:

No scalar data was found.

Probable causes:

You haven't written any scalar data to your event files.
TensorBoard can't find your event files.

If you're new to using TensorBoard, and want to find out how to add data and set up your event files, check out the README and perhaps the TensorBoard tutorial .

If you think TensorBoard is configured properly, please see the section of the README devoted to missing data problems and consider filing an issue on GitHub.

Yilong94

Yilong94 commented on May 8, 2017

@Yilong94

Facing the same issue as well. I am not sure if it is because of the logdir path such that it is unable to identify the Event file.

guschmue

guschmue commented on May 8, 2017

@guschmue
Contributor
MMCMA

MMCMA commented on May 8, 2017

@MMCMA
Author

no, not for me. The dir is not empty, I see the tfevents file in there, but once more the same error log in the browser.

Ipaulsen

Ipaulsen commented on May 10, 2017

@Ipaulsen

I am also seeing the same thing :(
Starting TensorBoard 47 at http://0.0.0.0:6006 (Press CTRL+C to quit) WARNING:tensorflow:path ../external/data/plugin/text/runs not found, sending 404 WARNING:tensorflow:path ../external/data/plugin/text/runs not found, sending 404 WARNING:tensorflow:path ../external/data/plugin/text/runs not found, sending 404 WARNING:tensorflow:path ../external/data/plugin/text/runs not found, sending 404
I see the tensorBoard at http://0.0.0.0:6006
no event data though

Ipaulsen

Ipaulsen commented on May 10, 2017

@Ipaulsen

Looks like this known issue it was fix in a later version. Got mine working by -
pip uninstall tensorflow
then
pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.11.0-py2-none-any.whl
This is for python 2.7

MMCMA

MMCMA commented on May 10, 2017

@MMCMA
Author

well, I am not sure if downgrading to tf.version <1.0.0 is the solution here.

Ipaulsen

Ipaulsen commented on May 11, 2017

@Ipaulsen

Your probably right cherry picking the patch is better until next release.

MMCMA

MMCMA commented on May 12, 2017

@MMCMA
Author

I just want this issue to be resolved, going forward it should be in the interest of everyone to achieve a more permanent fix. In the mean time #9854 has a fix for getting tensorboard working for version1.1.0.
tensorboard --logdir==training:model_dir --host=127.0.0.1

30 remaining items

apodemakeles

apodemakeles commented on Nov 23, 2017

@apodemakeles

same issue, tensorflow 1.4, tensorboard 0.4.0rc, windows 7, anaconda, py3.6
tensorboard --logdir==training:log_dir --host=127.0.0.1
works

ppaulggit

ppaulggit commented on Dec 17, 2017

@ppaulggit

Just update the tensorflow version or try to open tensorboard in your logdir
tensorboard --logdir=your_log_dir/your_speciafy_log_dir
or open tensorboard in your specific logdir
tensorboard --logdir=your_specific_log_dir
these two type can works, because the it depends on which path you open tensorboard

slideWXY

slideWXY commented on Dec 22, 2017

@slideWXY

replace " space " in your path location by renaming directories
That's work really well

Hope that will work for you guys 👍

alberduris

alberduris commented on Mar 14, 2018

@alberduris

I get this problem eventually. I don't know why sometimes it works and sometimes doesn't, but the work around is opening it with Firefox.

I can't really understand how a Google Product like Tensorboard doesn't work on Chrome and does it on Firefox, but that's it.

Same as #10611

Carolzheng

Carolzheng commented on Apr 19, 2018

@Carolzheng

Same issue. And it was working properly this morning until suddenly I get this error. I'm on Ubuntu 16.04 with python 3.6 and tensorflow 1.7.0, GPU version.
tensorboard --logdir==training:log_dir --host=127.0.0.1 worked for me. But absolutely no idea why it didn't work and why it worked.

toiara

toiara commented on Apr 22, 2018

@toiara

Faced the same issue. I'm on Win10 with python 3.6 and tensorflow 1.6 GPU.
I found the work around.
It is opening GRAPHS with Edge.
untitled

alaakh42

alaakh42 commented on Apr 30, 2018

@alaakh42

I was facing this problem on a Linux machine,
I discovered that this happened because I was in the directory of logs folder and by going back to the parent directory, and then issuing the command
tensorboard --logdir==training:path_to_logs/logs_folder_name --host=127.0.0.1 --port=6007
It worked fine.

diegovalenzuelaiturra

diegovalenzuelaiturra commented on Jul 10, 2018

@diegovalenzuelaiturra
kirankumarkv

kirankumarkv commented on Feb 4, 2019

@kirankumarkv

Had the same problem of graphs not populating with windows 10
tensorboard --logdir=training:log_dir --host=127.0.0.1

Hence specified just the directory containg training information
tensorboard --logdir=training --host=127.0.0.1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

    Development

    No branches or pull requests

      Participants

      @dartdog@aselle@teamdandelion@Ipaulsen@christabella

      Issue actions

        Tensorboard does not show any scalers or graphs · Issue #9701 · tensorflow/tensorflow