Simplify flags

README.md

@@ -99,7 +99,7 @@ python -m examples.router_chat --router mf --threshold 0.11593
 
 In the above examples, GPT-4 and Mixtral 8x7B are used as the model pair, but you can modify this using the `strong-model` and `weak-model` arguments.
 
-We leverage [LiteLLM](https://github.com/BerriAI/litellm) to support chat completions from a wide-range of open-source and closed models. In general, you need a setup an API key and point to the provider with the appropriate model name. Alternatively, you can also use **any OpenAI-compatible endpoint** by prefixing the model name with `openai/` using the `--alt-base-url` and `--alt-api-key` flags to point to the server.
+We leverage [LiteLLM](https://github.com/BerriAI/litellm) to support chat completions from a wide-range of open-source and closed models. In general, you need a setup an API key and point to the provider with the appropriate model name. Alternatively, you can also use **any OpenAI-compatible endpoint** by prefixing the model name with `openai/` and setting the `--base-url` and `--api-key` flags.
 
 Note that regardless of the model pair used, an `OPENAI_API_KEY` will be required to generate embeddings for both the `mf` and `sw_ranking` routers.
 

routellm/openai_server.py

@@ -39,8 +39,8 @@ async def lifespan(app):
         routers=args.routers,
         config=yaml.safe_load(open(args.config, "r")) if args.config else None,
         routed_pair=routed_pair,
-        alt_base_url=args.alt_base_url,
-        alt_api_key=args.alt_api_key,
+        api_base=args.base_url,
+        api_key=args.api_key,
         progress_bar=True,
     )
     yield
@@ -159,14 +159,14 @@ async def create_chat_completion(request: ChatCompletionRequest):
     choices=list(ROUTER_CLS.keys()),
 )
 parser.add_argument(
-    "--alt-base-url",
-    help="The base URL used for LLM requests",
+    "--base-url",
+    help="The base URL used for all LLM requests",
     type=str,
     default=None,
 )
 parser.add_argument(
-    "--alt-api-key",
-    help="The API key used for LLM requests",
+    "--api-key",
+    help="The API key used for all LLM requests",
     type=str,
     default=None,
 )