Tensorboard: Tensorboard does not work after updating Tensorflow to version 1.4

[ybsave]

I just updated my Tensorflow to version 1.4 with Tensorboard version 0.4.0rc2. However, the new version of Tensorboard cannot show anything for a log directory which was previously displayed without any problem before this update.

OS: Windows 10
Browser: Google Chrome Version 61.0.3163.100 (Official Build) (64-bit)
`Test results with "inspection" command:

Found event files in:
F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented
F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented\eval

These tags are in F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented:
audio -
histograms -
images
images/image/0
images/image/1
images/image/2
images/image/3
images/image/4
images/image/5
scalars
cross_entropy_1
global_step/sec
learning_rate_1
reg_loss
total_loss
train_accuracy_1
tensor -

Event statistics for F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented:
audio -
graph
first_step 0
last_step 0
max_step 0
min_step 0
num_steps 1
outoforder_steps []
histograms -
images
first_step 1
last_step 99901
max_step 99901
min_step 1
num_steps 1000
outoforder_steps []
scalars
first_step 1
last_step 99901
max_step 99901
min_step 1
num_steps 1000
outoforder_steps []
sessionlog:checkpoint
first_step 1
last_step 100000
max_step 100000
min_step 1
num_steps 100
outoforder_steps []
sessionlog:start
outoforder_steps []
steps [1, 4001, 8001, 12001, 16001, 20001, 24001, 28001, 32001, 36001, 40001, 44001, 48001, 52001, 56001, 60001, 64001, 68001, 72001, 76001, 80001, 84001, 88001, 92001, 96001]
sessionlog:stop -
tensor -

These tags are in F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented\eval:
audio -
histograms -
images -
scalars
accuracy
loss
tensor -

Event statistics for F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented\eval:
audio -
graph
first_step 0
last_step 0
max_step 0
min_step 0
num_steps 1
outoforder_steps []
histograms -
images -
scalars
first_step 4000
last_step 100000
max_step 100000
min_step 4000
num_steps 25
outoforder_steps []
sessionlog:checkpoint -
sessionlog:start -
sessionlog:stop -
tensor -`

[jart]

Does it log anything? Anything in the Chrome console log? Can you post a
screenshot?

On Nov 2, 2017 4:10 PM, "ybsave" [e-mail] wrote:

I just updated my Tensorflow to version 1.4 with Tensorboard version
0.4.0rc2. However, the new version of Tensorboard cannot show anything for
a log directory which was previously displayed without any problem before
this update.

OS: Windows 10
Browser: Google Chrome Version 61.0.3163.100 (Official Build) (64-bit)

Test results with "inspection" command:

Found event files in:
F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_
augmented
F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_
augmented\eval
These tags are in F:\projects\DeepLearning\TensorFlow\Models\Cifar\
densenet_40_12_cifar_10_augmented:
audio -
histograms -
images
images/image/0
images/image/1
images/image/2
images/image/3
images/image/4
images/image/5
scalars
cross_entropy_1
global_step/sec
learning_rate_1
reg_loss
total_loss
train_accuracy_1
tensor - Event statistics for F:\projects\DeepLearning\
TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented:
audio -
graph
first_step 0
last_step 0
max_step 0
min_step 0
num_steps 1
outoforder_steps []
histograms -
images
first_step 1
last_step 99901
max_step 99901
min_step 1
num_steps 1000
outoforder_steps []
scalars
first_step 1
last_step 99901
max_step 99901
min_step 1
num_steps 1000
outoforder_steps []
sessionlog:checkpoint
first_step 1
last_step 100000
max_step 100000
min_step 1
num_steps 100
outoforder_steps []
sessionlog:start
outoforder_steps []
steps [1, 4001, 8001, 12001, 16001, 20001, 24001, 28001, 32001, 36001,
40001, 44001, 48001, 52001, 56001, 60001, 64001, 68001, 72001, 76001,
80001, 84001, 88001, 92001, 96001]
sessionlog:stop -
tensor - These tags are in F:\projects\DeepLearning\
TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented\eval:
audio -
histograms -
images -
scalars
accuracy
loss
tensor - Event statistics for F:\projects\DeepLearning\
TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented\eval:
audio -
graph
first_step 0
last_step 0
max_step 0
min_step 0
num_steps 1
outoforder_steps []
histograms -
images -
scalars
first_step 4000
last_step 100000
max_step 100000
min_step 4000
num_steps 25
outoforder_steps []
sessionlog:checkpoint -
sessionlog:start -
sessionlog:stop -
tensor -

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/tensorflow/tensorboard/issues/695, or mute the thread
https://github.com/notifications/unsubscribe-auth/AADAbsFPPh1lBNT5kdaQS3VdGh8_kR-hks5sykvOgaJpZM4QQeEf
.

[ybsave]

it doesn't log anything. :( Just the regular message on the screen:

No dashboards are active for the current data set.

Probable causes:

You haven’t written any 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.

Last reload: Thu Nov 02 2017 16:36:02 GMT-0700 (Pacific Daylight Time)
Log directory: F:\projects\DeepLearning\TensorFlow\Models\Cifar\densenet_40_12_cifar_10_augmented

[ybsave]

I have downgraded my tensorboard to version 0.1.8. However, it still has the same problem. Does this mean Tensorflow 1.4 has changed some modules used by Tensorboard?

I uploaded a folder containing the summaries which were previously display without any problem before upgrading to TF 1.4. For your check.

Before upgrading to Tensorflow 1.4, I have no problem of using Tensorboard. Thank you.
my_resnet_32_version.zip

[ybsave]

Update: I downgraded my Tensorflow-GPU to version 1.3; it also downgraded the Tensorboard to version 0.1.8. However, the tensorboard still does not work. The log directory could be displayed correctly on other computers with Tensorflow-GPU 1.3 without updating to 1.4.

Does this mean that when upgrading to Tensorflow 1.4, the new installation program damaged some basic functions? Thus, even I downgrade Tensorflow again, Tensorboard still does not work!

Is there any viable solutions? Thank you.

[ybsave]

Update: I find out a solution. After I run tensorboard in the same drive, e.g., "F:\", everything works fine!

This seems a bug in Windows. In previous versions, I can run Tensorboard on other drives, or even across network without any problem. Hope it will be fixed soon. Thank you.

[jart]

I'm glad to hear you got things working. Thanks for reaching out and sharing this information with the community.

[pasikon]

Yeah, that was the issue for me as well, thanks for sharing

[obaidcuet]

@ybsave
That worked for me as well (start tensorboard from the same drive where logdir is)

[subhamiitk]

Yeah. I too ran it in the same directory and it worked. Thanks!