`BatchSizeFinder` with method `fit` and separate batch_size attributes for train and val (e.g., `self.train_batch_size` and `self.val_batch_size`)

Nobody has claimed this yet.

### Outline & Motivation

I suggest allowing `batch_arg_name` to accept a list of arg names. E.g. `tuner.scale_batch_size(..., batch_arg_name=["train_batch_size", "val_batch_size").

### Pitch

Fit uses both train and val dataloaders. They can have their own batch sizes.

### Additional context

_No response_

cc @lantiga @justusschock

Start by locating BatchSizeFinder and the scale_batch_size entry point, then trace how fit handles its train and validation dataloaders. Define the behavior for a list passed as batch_arg_name, including how separate train and validation batch sizes are tuned, and verify the existing batch-size tuning tests cover the new API.

Written by the indexing model from the issue text.