Fix whisper hook generate decorator arguments passing

[as-suvorov]

Description

Fix whisper hook generate decorator arguments passing
Transformer change introduced in: huggingface/transformers#35753 (4.54.1 release)
Tested with 4.53, 4.55 transformers

Ticket: 174921

Checklist:

• This patch fully addresses the ticket.

[Copilot]

Pull Request Overview

This PR fixes incorrect argument passing in the Whisper hook's generate decorator and corrects the use case class reference for speech-to-text models. The changes address a bug where the generate method wrapper was explicitly declaring an attention_mask parameter instead of accepting all keyword arguments dynamically.

Key Changes:

• Corrected the import and usage of UseCaseSpeech2Text instead of UseCaseTextToSpeech in speech-to-text model handling
• Fixed the my_generate wrapper function to properly forward all keyword arguments to the underlying generate method

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
tools/llm_bench/llm_bench_utils/ov_utils.py	Corrected import and class reference from UseCaseTextToSpeech to UseCaseSpeech2Text for proper speech-to-text model validation
tools/llm_bench/llm_bench_utils/hook_forward_whisper.py	Fixed my_generate wrapper to accept all keyword arguments instead of explicitly declaring attention_mask, ensuring proper argument forwarding

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[as-suvorov]

build_jenkins

[Wovchena]

I think the problem is in OV's Jenkins. We have a PR in a queue hanging for 20 hours. I'll contact @akladiev. But it's worth trying once more :)

[Wovchena]

build_jenkins