TensorBoard Tutorial

The tutorial consists of three parts:

• theoretical introduction - presentation
• examples - notebook
• exercises - notebook

Prerequisites

Uv

Verify if uv is installed in your environment, eg. by:

uv --version

If uv is not installed, you may install it with pip:

pip3 install uv

or with your system:

curl -LsSf https://astral.sh/uv/install.sh | sh

Set up

1. Set up the environment with:

./setup.sh

After this command, new Jupyter kernel (tensorboard-uv) is registered.

Use

1. Open tensorboard_examples.ipynb in Jupyter environment (eg. VS Code, Jupyter Lab)
2. Choose tensorboard-uv kernel in the notebook settings (top-right corner).

FAQ

No space left on device during set up

If gaining enough free space on the device is not possible, we recommend using this repository on an external device. In order to cache uv packages outside the default directory (~/.cache), use:

UV_CACHE_DIR="/path/to/new/cache/dir" ./setup.sh