Minor update

Author: yinggeh

README.md

@@ -110,11 +110,9 @@ In this case, please install vLLM first. You can do so by running
 container with the following commands:
 
 ```
-TEMP_TRITON_VLLM_DIR=/tmp/vllm_backend
-mkdir -p /opt/tritonserver/backends/vllm $TEMP_TRITON_VLLM_DIR
-git clone https://github.com/triton-inference-server/vllm_backend.git $TEMP_TRITON_VLLM_DIR
-cp -r $TEMP_TRITON_VLLM_DIR/src/* /opt/tritonserver/backends/vllm
-rm -rf $TEMP_TRITON_VLLM_DIR
+mkdir -p /opt/tritonserver/backends/vllm
+git clone https://github.com/triton-inference-server/vllm_backend.git /tmp/vllm_backend
+cp -r /tmp/vllm_backend/src/* /opt/tritonserver/backends/vllm
 ```
 
 ## Using the vLLM Backend

src/utils/metrics.py

@@ -83,7 +83,7 @@ def _log_counter(self, counter, data: Union[int, float]) -> None:
             counter.increment(data)
 
     def log(self, stats: VllmStats) -> None:
-        """Logs to triton metrics server every iteration.
+        """Report stats to Triton metrics server.
 
         Args:
             stats: Created by LLMEngine for use by VllmStatLogger.