[OV] Add quantization for text2text-generation models

[nikita-savelyevv]

What does this PR do?

Added post training quantization support for text2text-generation models.

Example metrics are below. Accuracy is computed with EleutherAI/lm-evaluation-harness#3101 on HellaSwag dataset.

Model ID	FP32 Accuracy	INT8 WC Accuracy	INT8 PTQ Accuracy	INT8 WC Speedup	INT8 PTQ Speedup
google-t5/t5-small	26.72%	26.78%	26.96%	1.06x	1.14x
google-t5/t5-large	34.84%	34.82%	34.80%	1.70x	1.81x
google-t5/t5-3b	38.52%	38.41%	38.79%	2.46x	2.38x

WC stands for weight compression, PTQ stands for Post-Training-Quantization.

Ticket 165770

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[echarlaix]

LGTM, thanks a lot @nikita-savelyevv !