split_between_processes for Dataset (#2433)
* split_between_processes for Dataset
* Update state.py
* remove param datasets.Dataset from split_between_processes, add note to function doc
* is_datasets_available is a function not a var
* reformat to make ruff happy
* isinstance(inputs, Dataset) only if is_datasets_available()
* add test_split_between_processes_dataset
* split_between_processes for Dataset: pad if apply_padding
* removed trailing whitespace
* complete test_split_between_processes_dataset
* fix test_split_between_processes_dataset for single GPU