Feature: support code completion (#3612)

python/sglang/srt/server_args.py

@@ -56,6 +56,7 @@ class ServerArgs:
     device: Optional[str] = None
     served_model_name: Optional[str] = None
     chat_template: Optional[str] = None
+    completion_template: Optional[str] = None
     is_embedding: bool = False
     revision: Optional[str] = None
 
@@ -456,6 +457,12 @@ class ServerArgs:
             default=ServerArgs.chat_template,
             help="The buliltin chat template name or the path of the chat template file. This is only used for OpenAI-compatible API server.",
         )
+        parser.add_argument(
+            "--completion-template",
+            type=str,
+            default=ServerArgs.completion_template,
+            help="The buliltin completion template name or the path of the completion template file. This is only used for OpenAI-compatible API server. only for code completion currently.",
+        )
         parser.add_argument(
             "--is-embedding",
             action="store_true",