[Serve LLM] `repetition_penality` missing from SamplingParams

[ArthurBook]

What happened + What you expected to happen

repetition_penality is a valid vLLM SamplingParam but is missing from the ray.serve counterpart. This causes it to be ignored when passed!

Versions / Dependencies

@latest

Reproduction script

from ray import serve
from ray.serve.llm import LLMConfig, LLMServer, LLMRouter

llm_config = LLMConfig(
    model_loading_config=dict(
        model_id="qwen-0.5b",
        model_source="Qwen/Qwen2.5-0.5B-Instruct",
    ),
    deployment_config=dict(
        autoscaling_config=dict(
            min_replicas=1, max_replicas=2,
        )
    ),
    # Pass the desired accelerator type (e.g. A10G, L4, etc.)
    accelerator_type="A10G",
    # You can customize the engine arguments (e.g. vLLM engine kwargs)
    engine_kwargs=dict(
        tensor_parallel_size=2,
    ),
)

# Deploy the application
deployment = LLMServer.as_deployment(llm_config.get_serve_options(name_prefix="vLLM:")).bind(llm_config)
llm_app = LLMRouter.as_deployment().bind([deployment])
serve.run(llm_app, blocking=False)

curl -X POST http://localhost:8000/v1/chat/completions \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer fake-key" \
     -d '{
           "model": "qwen-0.5b",
           "messages": [{"role": "user", "content": "Hello!"}]
           "repetition_penalty": 99,
         }'

Issue Severity

High: It blocks me from completing my task.