add sdpa and flash_attention2 support to speech2text

[avishaiElmakies]

I added support for sdpa and flash_attention2 to speech2text model (I started with sdpa but beacuse of "#copied" thought it would just be easier to add flash_attentnion2 as well). copy from bart.
addresses #26350 and #28005

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?

@amyeroberts @fxmarty

Notes:

1. there was a bug in the class Speech2TextSinusoidalPositionalEmbedding which caused test_eager_matches_sdpa_generate to fail. I fixed it. The class was copied in SpeechT5. so I fixed the bug there as well
2. the test test_flash_attn_2_generate_reuse_cache was failing because he wasn't meant for speech2text model. I copied and tweaked the test from whisper.
3. the only test currently failing on my machine is test_flash_attn_2_from_config with the error ValueError: Unrecognized configuration class <class 'transformers.models.speech_to_text.configuration_speech_to_text.Speech2TextConfig'> for this kind of AutoModel: AutoModelForCausalLM. . speech2text doesn't have a model for causalLM but i couldn't find what other models that have the same situation and what they do with the test. bart has BartForCausalLM. so would love some guidance here

[amyeroberts]

cc @ylacombe for first review