I am currently using Tensorflow on Linux, and when I launch tensorboard and go to the localhost using Firefox, I am only able to see what's enclosed in the image. I am unable to interact with anything, and the whole thing is unresponsive. How do you think I might be able to fix this?
Tensorboard Window
Related
my colab restart after showing memory error.
I am trying to train cnn model on google colab notebook but after fist epoch my notebook restart itself by showing the error of memory allocation issue.
I was running the same one day before and it was running fine.
Logs are attached here.
It sounds like you might be running out of memory in google Colab.
One way to get around this might be to use your own computer. The method for this with anaconda is:
open anaconda
type in to anaconda: pip install jupyter
Once it is downloaded, type: jupyter notebook
this then opens a webpage, in the webpage there is a dropdown (I think called new) which you should click on, then click on terminal
in the terminal, you should type in (all on one line): jupyter notebook --NotebookApp.allow_origin='https://colab.research.google.com' --port=8889 --NotebookApp.port_retries=0
There will then be 3 output links, copy the second one.
back in colab, click the dropdown in the top right next to connect, then click connect to local runtime
Then input the copied URL into the prompt, then connect.
I am connecting to local run time in Google colab. I tried loading tensorboard. It does not give any error but I can only see a blank space. It works fine if I connect to hosted runtime in colab.
You can refer to this article and colab to setup tensorboard in Google Colab.
In some cases, the logdir you're pointing tensorboard to probably does not exist, Please check and make sure you are pointing to correct log directory. Thank you!
I want to know the state of the gpu used by mine when using Colab.
I tried
!nivdia-smi
while it can not work along with my program so It shows no process.
I never tried but there are ways to ssh into colab so you can run nvidia-smi on ssh and run your code on regular colab cell.
Click on the menu "Runtime" >> "Manage Session".
I think there's the information you need.
change
!nivdia-smi
to
!nvidia-smi
I am running the latest version of tensorflow (1.7.0) in Windows. It added a new feature for debugging in tensorboard. However, I am not able to get it to connect. I was previously using the LocalCLIDebugWrapperSession debugger. However, I wanted to play with the new features in TensorBoardDebugWrapperSession.
When I run tensorboard using the command "tensorboard --logdir=dir --debugger_port 6064" it tells me "Creating InteractiveDebuggerPlugin at port 6064".
However, when I run my model in tensorflow (after swapping out the LocalCLIDebugWrapperSession for the TensorBoardDebugWrapperSession) it never seems to connect to the interactive debugger. I don't get an error in the output for either tensorboard or tensorflow. Nothing happens and on the tensorboard debugger page it just shows: "Debugger is waiting for Session.run() connections..."
It seems that tensorflow does not have support for grpc connection on windows. A workaround is to use windows subsystem. Note that there is no direct access to system GPU from the subsystem. So it cannot be use with tensorflow-gpu version
start_epoch: grpc:// debug URL scheme is not implemented on Windows yet.
https://github.com/tensorflow/tensorflow/issues/17933
Do I absolutely need to use jupyter notebook to run TensorFlow in Windows ?
I tried the detect object example with the jupyter notebook, it works but I'm not really comfortable, Im used to notepad++ and running python directly on my windows without virtual environment.
I tried to copy past all the codes but I run into many hugs.
No, it is not compulsory to use Jupyter notebook to run Tensorflow on Windows. I personally use PyCharm as my IDE and Anaconda for dependency management (this is completely optional).
I would recommend you to use a proper IDE instead of notepad++ because it's much easier to do debugging using an IDE. You'll also be cloning a lot from Git when you start developing your own model, and usually the open source models out there has a lot of classes and methods in it (take Google's Inception net for example).
Another alternative would be maybe you can start posting about the bugs you are facing, then we can all start helping you.