Training Visualization

## Description
DJL should include a NN web or GUI visualization tool like [DL4J](https://deeplearning4j.org/docs/latest/deeplearning4j-nn-visualization) to help optimize NN parameters and NN layers.

### Will this change the current api? How?
This will be a new API.

### Who will benefit from this feature?
Users who want to monitor training jobs and help determine why their model is not training successfully.

## References
- See also: [Tensorboard](https://www.tensorflow.org/tensorboard)

Start by reviewing DJL's existing API structure and the linked DL4J neural-network visualization and TensorBoard references. Define the scope of the proposed web or GUI tool, including how it monitors training jobs and exposes model parameters and layers. Done means a documented design and a working visualization API integrated with DJL.

Written by the indexing model from the issue text.