Bump the version to v0.1.8 (#93)

python/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sglang"
-version = "0.1.7"
+version = "0.1.8"
 description = "A structured generation langauge for LLMs." 
 readme = "README.md"
 requires-python = ">=3.8"

python/sglang/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.7"
+__version__ = "0.1.8"
 
 from sglang.api import *
 from sglang.global_config import global_config

python/sglang/srt/managers/tokenizer_manager.py

@@ -50,7 +50,9 @@ def init_global_processor(server_args: ServerArgs):
     )
 
 
-def get_pixel_values(image_data, image_aspect_ratio=None, image_grid_pinpoints=None, processor=None):
+def get_pixel_values(
+    image_data, image_aspect_ratio=None, image_grid_pinpoints=None, processor=None
+):
     try:
         processor = processor or global_processor
         image = load_image(image_data)
@@ -117,14 +119,22 @@ class TokenizerManager:
 
     async def get_pixel_values(self, image_data):
         aspect_ratio = getattr(self.hf_config, "image_aspect_ratio", None)
-        grid_pinpoints = self.hf_config.image_grid_pinpoints if aspect_ratio == "anyres" else None
+        grid_pinpoints = (
+            self.hf_config.image_grid_pinpoints if aspect_ratio == "anyres" else None
+        )
         if self.executor is not None:
             loop = asyncio.get_event_loop()
             return await loop.run_in_executor(
-                self.executor, get_pixel_values, image_data, aspect_ratio, grid_pinpoints
+                self.executor,
+                get_pixel_values,
+                image_data,
+                aspect_ratio,
+                grid_pinpoints,
             )
         else:
-            return get_pixel_values(image_data, aspect_ratio, grid_pinpoints, self.processor)
+            return get_pixel_values(
+                image_data, aspect_ratio, grid_pinpoints, self.processor
+            )
 
     async def generate_request(self, obj: GenerateReqInput):
         if self.to_create_loop:

python/sglang/srt/server.py

@@ -380,12 +380,13 @@ def launch_server(server_args, pipe_finish_writer):
 
         success = False
         for i in range(60):
+            time.sleep(1)
             try:
                 res = requests.get(url + "/get_model_info", timeout=5)
                 success = True
                 break
             except requests.exceptions.RequestException as e:
-                time.sleep(1)
+                pass
 
         if success:
             pipe_finish_writer.send("init ok")
@@ -406,7 +407,7 @@ class Runtime:
         model_mode: List[str] = (),
         schedule_heuristic: str = "lpm",
         random_seed: int = 42,
-        log_level: str = "warning",
+        log_level: str = "error",
     ):
         host = "127.0.0.1"
         port = alloc_usable_network_port(1)[0]