[Feature]: Support for serving two llms at the same time

[czg1225]

🚀 The feature, motivation and pitch

I want to implement the interaction between two llms. But I found it will encounter error or lock when initializing two llms like following. Could you help to support this feature? It will be great useful!

first_model_name = "Qwen/Qwen2.5-1.5B-Instruct"
second_model_name = "Qwen/Qwen2.5-7B-Instruct"
first_llm = LLM(model=first_model_name, tensor_parallel_size=2)
second_llm = LLM(model=second_model_name, tensor_parallel_size=2)

Alternatives

No response

Additional context

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.