# Triton Backend for vLLM.
backend: "vllm"
platform: "vllm"
max_batch_size: 0

input [
]

output [
]




instance_group [
  {
    kind: KIND_AUTO
    count: 1
  }
]