add parameter

entrypoint.sh

@@ -25,16 +25,26 @@ if [[ ! -z "${MAX_MODEL_LEN}" ]]; then
     additional_args="${additional_args} --max-model-len ${MAX_MODEL_LEN}"
 fi
 
-if [[ ! -z "${ROOT_PATH}" ]]; then
-    additional_args="${additional_args} --root-path ${ROOT_PATH}"
+if [[ ! -z "${TENSOR_PARALLEL_SIZE}" ]]; then
+    additional_args="${additional_args} --tensor-parallel-size ${TENSOR_PARALLEL_SIZE}"
 fi
 
-# Run the provided command
-# exec python3 -u -m vllm.entrypoints.openai.api_server \
-# --model "${HF_MODEL}" \
-# --host 0.0.0.0 \
-# --port 7860 \
-# ${additional_args}
+if [[ ! -z "${DOWNLAD_DIR}" ]]; then
+    additional_args="${additional_args} --download-dir ${DOWNLAD_DIR}"
+fi
+
+if [[ ! -z "${ENFORCE_EAGER}" ]]; then
+    additional_args="${additional_args} --enforce-eager"
+fi
+
+if [[ ! -z "${SERVED_MODEL_NAME}" ]]; then
+    additional_args="${additional_args} --served-model-name ${SERVED_MODEL_NAME}"
+fi
+
+# PATH_MODEL="/data/models--${HF_MODEL/\//--}"
+# if [ -d "$PATH_MODEL" ]; then
+#   HF_MODEL=$PATH_MODEL
+# fi
 
 exec python3 -u api_server.py \
     --model "${HF_MODEL}" \