What does this PR do?

This PR is following the #3362 to give the freedom for user to be able to pad_to_multiple_of. It is tested using simple following script:

from datasets import load_dataset
from trl import SFTTrainer, SFTConfig


dataset = load_dataset("roneneldan/TinyStories", split="train[:1%]") 

config = SFTConfig(
    output_dir="sft_pad_multiple_of_2",
    max_steps=10, 
    pad_to_multiple_of=4,
    logging_steps=1,
    report_to=[],
)


if __name__ == "__main__":
    
    trainer = SFTTrainer(
        model="Qwen/Qwen2-0.5B-Instruct",
        train_dataset=dataset,
        args=config)
    trainer.train()

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a GitHub issue? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.