Support relative paths in ModelCheckpoint state

Nobody has claimed this yet.

## 🚀 Feature

It would be great if `ModelCheckpoint` internal state supported relative paths.

### Motivation

Currently, if you specify a relative path for `dirpath`, `ModelCheckpoint` converts it to an absolute path under the hood. This makes it hard to resume training if the log directory is moved, or if resuming training from a different server with a different directory structure.

For example, I specify the relative path `a/b/c`, and `ModelCheckpoint` converts it to `cwd()/a/b/c`. The model trains correctly for a while. Then, HTCondor reschedules my job on a different server. Now, the `cwd()` is different, even though the relative path `a/b/c` is still the same. The job is unable to resume from checkpoint and I get:

> UserWarning: The dirpath has changed from to , therefore `best_model_score`, `kth_best_model_path`, `kth_value`, `last_model_path` and `best_k_models` won’t be reloaded.

### Pitch

Create an argument `relative_paths=True` that would allow `ModelCheckpoint` to use relative paths in its internal state.

### Alternatives

User can create their own checkpoint callback that supports relative paths. But, it would be much nicer if Lightning supported it :-)

### Additional context

______________________________________________________________________

#### If you enjoy Lightning, check out our other projects! ⚡

- [**Metrics**](https://github.com/Lightning-AI/metrics): Machine learning metrics for distributed, scalable PyTorch applications.

- [**Lite**](https://pytorch-lightning.readthedocs.io/en/latest/starter/lightning_lite.html): enables pure PyTorch users to scale their existing code on any kind of device while retaining full control over their own loops and optimization logic.

- [**Flash**](https://github.com/Lightning-AI/lightning-flash): The fastest way to get a Lightning baseline! A collection of tasks for fast prototyping, baselining, fine-tuning, and solving problems with deep learning.

- [**Bolts**](https://github.com/Lightning-AI/lightning-bolts): Pretrained SOTA Deep Learning models, callbacks, and more for research and production with PyTorch Lightning and PyTorch.

- [**Lightning Transformers**](https://github.com/Lightning-AI/lightning-transformers): Flexible interface for high-performance research using SOTA Transformers leveraging PyTorch Lightning, Transformers, and Hydra.

cc @borda @carmocca @awaelchli @ninginthecloud @jjenniferdai @rohitgr7 @akihironitta

Start by locating the Python ModelCheckpoint callback and the tests covering checkpoint state restoration. Reproduce a resume with a relative dirpath after changing the working directory, then verify that the internal paths and saved state restore correctly without the directory-change warning.

Written by the indexing model from the issue text.