vllm_info = """
Important: This endpoint is running vllm-gptq and not all Oobabooga parameters are supported.