tensorboard_logger

The `tensorboard_logger` library enables users to log numerical events to TensorBoard without the need for TensorFlow, facilitating the visualization of training progress and variable comparisons across different runs. Key features include the ability to specify logging directories for each run, log values associated with specific steps to maintain order, and an easy setup process that allows integration with standard TensorBoard functionalities. The library is particularly useful for projects not utilizing TensorFlow but still requiring tensor-based logging and visualization capabilities.