After reviewing some online tutorials regarding Tensorboard, I've tried implementing it in my research notebook but I can't seem to get this to work.

I've set

%load_ext tensorboard

near the top of my notebook, but the callback in my model doesn't seem to trigger anything. I would expect a web page to pop up like what running the CLI command does:

lean research “My Notebook"

From what I've read, Tensorboard runs typically on port 6006. I tried installing a separate docker image following the instructions - it failed due to “ports already in use”, so I converted the command to:

docker run -it -p 6006:8888 tensorflow/tensorflow:nightly-py3-jupyter

Given the docker image script seems to want to expose 8888 internally, I mapped 6006 to it and it loads some sort of gateway UI that is asking for login credentials for some reason. I'm not at all familiar with Tensorboard or even if I've set things up correctly. Has anyone else successfully used Tensorboard with Lean CLI yet?