tensorboard: tensorflow 1.8.0 has requirement tensorboard<1.9.0,>=1.8.0, but you'll have tensorboard 1.6.0 which is incompatible.

The tensorboard pip package is now just tensorboard, you should pip uninstall tensorflow-tensorboard and then pip install --upgrade tensorflow by itself (which will automatically install the correct version of tensorboard).

@wxprofessor It looks like your pip install tensorflow command is saying that you also have tensorflow-gpu installed - you want to only have one of the two installed at a time. Can you try uninstalling all the tensor* packages and then reinstalling? Or even better, get a fresh conda environment and try just installing tensorflow-gpu?

hopefully this will help? 😃

After reading through some blogs about the tensorboard issue, I just uninstalled tensorflow, and protobuf (because of a previous incompatibility), and reinstalled tensorflow hoping the new dependencies would be found, and received the following.

I then restarted the kernel in Jupyter Lab and ran a simple tensorflow gpu test to receive the original error I was trying to fix 😃

I found the potential fix for that here in the boards: https://github.com/tensorflow/tensorflow/issues/20778

that stated uninstalling and reinstalling might be a fix 😃

Hopefully this might help to point to a potential resolution, as I repeatedly beat my head against the wall 😃

You’ll want to install a more recent version of tensorflow-gpu than 1.5.0: https://pypi.org/project/tensorflow-gpu/#history

I don’t know what tensorboard_logger is - that’s not a module that we provide, so you’d have to ask whoever provides that module for support.

Thanks for the support @nfelt …its solved