sglang 0.1.20__py3-none-any.whl → 0.1.22__py3-none-any.whl

sglang/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.20"
+__version__ = "0.1.22"
 
 # SGL API Components
 from sglang.api import (
@@ -22,16 +22,16 @@ from sglang.api import (
     video,
 )
 
-# SGL Backends
-from sglang.backend.anthropic import Anthropic
-from sglang.backend.litellm import LiteLLM
-from sglang.backend.openai import OpenAI
-from sglang.backend.runtime_endpoint import RuntimeEndpoint
-from sglang.backend.vertexai import VertexAI
-
 # Global Configurations
 from sglang.global_config import global_config
 
+# SGL Backends
+from sglang.lang.backend.anthropic import Anthropic
+from sglang.lang.backend.litellm import LiteLLM
+from sglang.lang.backend.openai import OpenAI
+from sglang.lang.backend.runtime_endpoint import RuntimeEndpoint
+from sglang.lang.backend.vertexai import VertexAI
+
 # public APIs management
 __all__ = [
     "global_config",

sglang/api.py

@@ -4,8 +4,8 @@ import os
 import re
 from typing import Callable, List, Optional, Union
 
-from sglang.backend.base_backend import BaseBackend
 from sglang.global_config import global_config
+from sglang.lang.backend.base_backend import BaseBackend
 from sglang.lang.ir import (
     SglExpr,
     SglExprList,

sglang/backend/runtime_endpoint.py

@@ -12,7 +12,6 @@ from sglang.utils import http_request
 
 
 class RuntimeEndpoint(BaseBackend):
-
     def __init__(
         self,
         base_url: str,
@@ -38,7 +37,8 @@ class RuntimeEndpoint(BaseBackend):
         self.model_info = res.json()
 
         self.chat_template = get_chat_template_by_model_path(
-            self.model_info["model_path"])
+            self.model_info["model_path"]
+        )
 
     def get_model_name(self):
         return self.model_info["model_path"]
@@ -124,7 +124,12 @@ class RuntimeEndpoint(BaseBackend):
         else:
             raise RuntimeError(f"Invalid dtype: {sampling_params.dtype}")
 
-        for item in ["return_logprob", "logprob_start_len", "top_logprobs_num", "return_text_in_logprobs"]:
+        for item in [
+            "return_logprob",
+            "logprob_start_len",
+            "top_logprobs_num",
+            "return_text_in_logprobs",
+        ]:
             value = getattr(sampling_params, item, None)
             if value is not None:
                 data[item] = value
@@ -171,7 +176,12 @@ class RuntimeEndpoint(BaseBackend):
         else:
             raise RuntimeError(f"Invalid dtype: {sampling_params.dtype}")
 
-        for item in ["return_logprob", "logprob_start_len", "top_logprobs_num", "return_text_in_logprobs"]:
+        for item in [
+            "return_logprob",
+            "logprob_start_len",
+            "top_logprobs_num",
+            "return_text_in_logprobs",
+        ]:
             value = getattr(sampling_params, item, None)
             if value is not None:
                 data[item] = value

sglang/backend/vertexai.py

@@ -1,8 +1,6 @@
 import os
 import warnings
-from typing import List, Optional, Union
-
-import numpy as np
+from typing import Optional
 
 from sglang.backend.base_backend import BaseBackend
 from sglang.lang.chat_template import get_chat_template
@@ -21,7 +19,7 @@ except ImportError as e:
 
 
 class VertexAI(BaseBackend):
-    def __init__(self, model_name):
+    def __init__(self, model_name, safety_settings=None):
         super().__init__()
 
         if isinstance(GenerativeModel, Exception):
@@ -33,6 +31,7 @@ class VertexAI(BaseBackend):
 
         self.model_name = model_name
         self.chat_template = get_chat_template("default")
+        self.safety_settings = safety_settings
 
     def get_chat_template(self):
         return self.chat_template
@@ -54,6 +53,7 @@ class VertexAI(BaseBackend):
         ret = GenerativeModel(self.model_name).generate_content(
             prompt,
             generation_config=GenerationConfig(**sampling_params.to_vertexai_kwargs()),
+            safety_settings=self.safety_settings,
         )
 
         comp = ret.text
@@ -78,6 +78,7 @@ class VertexAI(BaseBackend):
             prompt,
             stream=True,
             generation_config=GenerationConfig(**sampling_params.to_vertexai_kwargs()),
+            safety_settings=self.safety_settings,
         )
         for ret in generator:
             yield ret.text, {}