Reputation: 43
Im running a pod in Jupyterlab from a Docker image. The Docker image is based on the official tensorflow/tensorflow:1.15.0-py3
image, and I installed every depenencies as described here.
Whenever I try to launch Tensorboard, I get:
'Launcher Error Invalid response: 500 Internal Server Error'
Any idea on this? Im on Jupyter notebook 6.0.2 Everything else works fine.
Thanks
Upvotes: 2
Views: 1970
Reputation: 122
I encountered the same issue. Make sure your run the tensorboard commands in the System terminal (with the sagemaker-user@studio
user) and not in your Kernel/Notebook terminal.
Upvotes: 0
Reputation: 1126
Your problem is mentioned in this github issue https://github.com/chaoleili/jupyterlab_tensorboard/issues/25.
This is likely due to some recent refactoring of Tensorflow. More specifically, the removal of TensorBoardWSGI
. See https://github.com/tensorflow/tensorboard/pull/3643.
Currently, the easiest way to fix it will be to use a lower version of Tensorboard. The latest version where TensorBoardWSGI
is available is 2.2.
pip install tensorboard==2.2
Should fix your problem.
Upvotes: 2