tiny remove deprecated endpoint call (#13607)

python/sglang/bench_one_batch_server.py

@@ -339,7 +339,7 @@ def run_one_case(
     output_throughput = batch_size * output_len / (latency - last_ttft)
     overall_throughput = batch_size * (input_len + output_len) / latency
 
-    server_info = requests.get(url + "/get_server_info").json()
+    server_info = requests.get(url + "/server_info").json()
     internal_state = server_info.get("internal_states", [{}])
     last_gen_throughput = internal_state[0].get("last_gen_throughput", None) or -1
     acc_length = internal_state[0].get("avg_spec_accept_length", None) or -1
@@ -451,7 +451,7 @@ def run_benchmark(server_args: ServerArgs, bench_args: BenchArgs):
         proc, base_url = launch_server_process(server_args)
 
     # Get tokenizer
-    server_info = requests.get(base_url + "/get_server_info").json()
+    server_info = requests.get(base_url + "/server_info").json()
     if "tokenizer_path" in server_info:
         tokenizer_path = server_info["tokenizer_path"]
     elif "prefill" in server_info:

python/sglang/bench_serving.py

@@ -2062,7 +2062,7 @@ async def benchmark(
 
     if "sglang" in backend:
         server_info = requests.get(
-            base_url + "/get_server_info", headers=get_auth_headers()
+            base_url + "/server_info", headers=get_auth_headers()
         )
         if server_info.status_code == 200:
             server_info_json = server_info.json()
@@ -2180,7 +2180,7 @@ async def benchmark(
     print("{:<40} {:<10.2f}".format("Max ITL (ms):", metrics.max_itl_ms))
     print("=" * 50)
 
-    resp = requests.get(base_url + "/get_server_info", headers=get_auth_headers())
+    resp = requests.get(base_url + "/server_info", headers=get_auth_headers())
     server_info = resp.json() if resp.status_code == 200 else None
 
     if (

python/sglang/lang/backend/runtime_endpoint.py

@@ -38,7 +38,7 @@ class RuntimeEndpoint(BaseBackend):
         self.verify = verify
 
         res = http_request(
-            self.base_url + "/get_model_info",
+            self.base_url + "/model_info",
             api_key=self.api_key,
             verify=self.verify,
         )
@@ -66,7 +66,7 @@ class RuntimeEndpoint(BaseBackend):
 
     def get_server_info(self):
         res = http_request(
-            self.base_url + "/get_server_info",
+            self.base_url + "/server_info",
             api_key=self.api_key,
             verify=self.verify,
         )
@@ -514,7 +514,7 @@ class Runtime:
 
     async def get_server_info(self):
         async with aiohttp.ClientSession() as session:
-            async with session.get(f"{self.url}/get_server_info") as response:
+            async with session.get(f"{self.url}/server_info") as response:
                 if response.status == 200:
                     return await response.json()
                 else:

python/sglang/profiler.py

@@ -41,7 +41,7 @@ def run_profile(
     # Dump server args.
     file_path = Path(output_dir) / "server_args.json"
     if not file_path.exists():
-        response = requests.get(url + "/get_server_info")
+        response = requests.get(url + "/server_info")
         response.raise_for_status()
         server_args_data = response.json()
         with open(file_path, "w") as file:

python/sglang/srt/entrypoints/http_server.py

@@ -1480,7 +1480,7 @@ def _execute_server_warmup(
     for _ in range(120):
         time.sleep(1)
         try:
-            res = requests.get(url + "/get_model_info", timeout=5, headers=headers)
+            res = requests.get(url + "/model_info", timeout=5, headers=headers)
             assert res.status_code == 200, f"{res=}, {res.text=}"
             success = True
             break

python/sglang/srt/server_args.py

@@ -2797,7 +2797,7 @@ class ServerArgs:
         parser.add_argument(
             "--preferred-sampling-params",
             type=str,
-            help="json-formatted sampling settings that will be returned in /get_model_info",
+            help="json-formatted sampling settings that will be returned in /model_info",
         )
 
         # LoRA