sglang 0.4.2.post3__tar.gz → 0.4.3__tar.gz

{sglang-0.4.2.post3/sglang.egg-info → sglang-0.4.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: sglang
-Version: 0.4.2.post3
+Version: 0.4.3
 Summary: SGLang is yet another fast serving framework for large language models and vision language models.
 License:                                  Apache License
                                    Version 2.0, January 2004
@@ -236,14 +236,15 @@ Requires-Dist: torchao>=0.7.0; extra == "runtime-common"
 Requires-Dist: uvicorn; extra == "runtime-common"
 Requires-Dist: uvloop; extra == "runtime-common"
 Requires-Dist: xgrammar>=0.1.10; extra == "runtime-common"
+Requires-Dist: ninja; extra == "runtime-common"
 Provides-Extra: srt
 Requires-Dist: sglang[runtime_common]; extra == "srt"
 Requires-Dist: cuda-python; extra == "srt"
-Requires-Dist: sgl-kernel>=0.0.3.post2; extra == "srt"
+Requires-Dist: sgl-kernel>=0.0.3.post6; extra == "srt"
 Requires-Dist: torch; extra == "srt"
-Requires-Dist: vllm==0.6.4.post1; extra == "srt"
-Requires-Dist: flashinfer_python>=0.2.0.post2; extra == "srt"
-Requires-Dist: outlines<0.1.0,>=0.0.44; extra == "srt"
+Requires-Dist: vllm<=0.7.2,>=0.6.4.post1; extra == "srt"
+Requires-Dist: flashinfer_python>=0.2.1.post1; extra == "srt"
+Requires-Dist: outlines<=0.1.11,>=0.0.44; extra == "srt"
 Provides-Extra: srt-hip
 Requires-Dist: sglang[runtime_common]; extra == "srt-hip"
 Requires-Dist: torch; extra == "srt-hip"

{sglang-0.4.2.post3 → sglang-0.4.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sglang"
-version = "0.4.2.post3"
+version = "0.4.3"
 description = "SGLang is yet another fast serving framework for large language models and vision language models."
 readme = "README.md"
 requires-python = ">=3.8"
@@ -21,12 +21,13 @@ runtime_common = [
     "hf_transfer", "huggingface_hub", "interegular", "modelscope",
     "orjson", "packaging", "pillow", "prometheus-client>=0.20.0",
     "psutil", "pydantic", "python-multipart", "pyzmq>=25.1.2",
-    "torchao>=0.7.0", "uvicorn", "uvloop", "xgrammar>=0.1.10"
+    "torchao>=0.7.0", "uvicorn", "uvloop", "xgrammar>=0.1.10", "ninja"
 ]
 srt = [
     "sglang[runtime_common]", "cuda-python",
-    "sgl-kernel>=0.0.3.post2", "torch", "vllm==0.6.4.post1",
-    "flashinfer_python>=0.2.0.post2", "outlines>=0.0.44,<0.1.0"
+    "sgl-kernel>=0.0.3.post6", "torch", "vllm>=0.6.4.post1,<=0.7.2",
+    "flashinfer_python>=0.2.1.post1",
+    "outlines>=0.0.44,<=0.1.11",
 ]
 
 # HIP (Heterogeneous-computing Interface for Portability) for AMD

{sglang-0.4.2.post3 → sglang-0.4.3}/sglang/check_env.py

@@ -19,6 +19,7 @@ def is_cuda_v2():
 # List of packages to check versions
 PACKAGE_LIST = [
     "sglang",
+    "sgl_kernel",
     "flashinfer",
     "triton",
     "transformers",

{sglang-0.4.2.post3 → sglang-0.4.3}/sglang/global_config.py

@@ -38,5 +38,7 @@ class GlobalConfig:
         self.enable_precache_with_tracing = True
         self.enable_parallel_encoding = True
 
+        self.enable_flashinfer_mla = False
+
 
 global_config = GlobalConfig()

{sglang-0.4.2.post3 → sglang-0.4.3}/sglang/srt/constrained/outlines_backend.py

@@ -35,7 +35,10 @@ is_hip_ = is_hip()
 if is_hip_:
     from outlines_core.fsm.json_schema import build_regex_from_schema
 else:
-    from outlines.fsm.json_schema import build_regex_from_schema
+    try:
+        from outlines.fsm.json_schema import build_regex_from_schema
+    except ImportError:
+        from outlines_core.fsm.json_schema import build_regex_from_schema
 
 
 logger = logging.getLogger(__name__)

{sglang-0.4.2.post3 → sglang-0.4.3}/sglang/srt/entrypoints/engine.py

@@ -297,7 +297,7 @@ def _set_envs_and_config(server_args: ServerArgs):
     # Set global environments
     os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
     os.environ["NCCL_CUMEM_ENABLE"] = "0"
-    os.environ["NCCL_NVLS_ENABLE"] = "0"
+    os.environ["NCCL_NVLS_ENABLE"] = str(int(server_args.enable_nccl_nvls))
     os.environ["TORCH_NCCL_AVOID_RECORD_STREAMS"] = "1"
     os.environ["CUDA_DEVICE_MAX_CONNECTIONS"] = "4"
 
@@ -317,7 +317,7 @@ def _set_envs_and_config(server_args: ServerArgs):
     if server_args.attention_backend == "flashinfer":
         assert_pkg_version(
             "flashinfer_python",
-            "0.2.0.post2",
+            "0.2.1.post1",
             "Please uninstall the old version and "
             "reinstall the latest version by following the instructions "
             "at https://docs.flashinfer.ai/installation.html.",