Feat/gradient statistics monitor callback

[JanFidor]

What does this PR do?

Closes #21589

Credits to @ZiadiSafouene whose original PR #21596 this one was heavily based on.

This PR provides implementation of GradientStatsMonitor callback for Lightning that tracks gradient statistics
during training.

The callbacks allows the user to track both global and per layer gradient statistics as well as per epoch and per batch ones.
Implementation should make extending this class simple with functions like compute_batch_stats or per epoch statistics functions following conventions from torchmetrics: initialize, update, compute.
The callback supports state_dicts -> interrupting and resuming training.
Provides easy diagnostics for high gradients even when working with gradient clipping by utilizing on_before_optimizer_step hook.
Related tests are included.

In short this callback provides a simple out of the box solution for tracking ML model gradients.

Before submitting

• Was this discussed/agreed via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

Reviewer checklist

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

---

📚 Documentation preview 📚: https://pytorch-lightning--21720.org.readthedocs.build/en/21720/