Adding a tutorial/example on how to track and visualize model metrics and performance

## 📚 Documentation

I would like to add a new example to the documents on how users could use tools like Tensorboard and Weights & Biases to track and visualize the performance of their models built in DGL. This would include how users can keep track of different experiments, compare model inputs and corresponding outputs.

This example would include
- A section on logging model performance metrics to tensorboard.
- A section on logging model performance metrics along with model inputs and outputs to W&B.

Please let me know what you guys think!