Support Kokoro model

[openvino-agent]

What does this PR do?

Set-up environment:

pip install git+https://github.com/openvino-agent/optimum-intel.git@support_kokoro
pip install kokoro

Run exporting the model:

optimum-cli export openvino -m hexgrad/Kokoro-82M ov_Kokoro-82M --trust-remote-code

Run inference:

import soundfile as sf
import torch
from optimum.intel import OVModelForTextToSpeechSeq2Seq

model_id = "hexgrad/Kokoro-82M"

# Export from Hub and save locally
ov_pipe = OVModelForTextToSpeechSeq2Seq.from_pretrained(model_id, trust_remote_code=True)

# Preprocess text into model inputs (G2P + tokenization + voice loading)
inputs = ov_pipe.preprocess_input(
    text="Hello, this is a test of Kokoro text-to-speech using OpenVINO.",
    voice="af_heart",
    speed=1.0,
)

# Generate audio waveform
speech = ov_pipe.generate(**inputs)

sf.write("speech.wav", speech.numpy(), samplerate=24000)

Fixes # (issue)

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?