diff --git a/compose-vLLM.yaml b/compose-vLLM.yaml index ebabfd6..d71ecc4 100644 --- a/compose-vLLM.yaml +++ b/compose-vLLM.yaml @@ -29,7 +29,7 @@ services: - "--port" - "8000" - "--gpu-memory-utilization" - - "0.85" + - "0.72" - "--max-model-len" - "65536" - "--hf-overrides"