[VITS] Add speaking_rate as an optionl forward argument

[gau-nernst]

What does this PR do?

Currently, to adjust the speed of generated audio with VitsModel, I have to modify the .speaking_rate attribute

from transformers import VitsModel, AutoTokenizer
import torch
from IPython.display import Audio

model_id = "facebook/mms-tts-eng"
model = VitsModel.from_pretrained(model_id)
tokenizer = AutoTokenizer.from_pretrained(model_id)

text = "some example text in the English language"
inputs = tokenizer(text, return_tensors="pt")

with torch.no_grad():
    model.speaking_rate = 2.0
    output = model(**inputs).waveform

Audio(output.numpy(), rate=model.config.sampling_rate)

To make it more convenient, I added speaking_rate as a forward argument, so speaking rate can be adjusted per-forward basis.

with torch.no_grad():
    output = model(**inputs, speaking_rate=0.5).waveform

The default behavior is preserved - if speaking_rate is not specified, it fallbacks to self.speaking_rate

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 or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• 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.

@eustlb

[Rocketknight1]

cc @eustlb @ebezzam

[gau-nernst]

@Rocketknight1 hi, any updates? thanks

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: vits