Fix multimodal registry and code sync scripts (#10759)

Co-authored-by: cctry <shiyang@x.ai>

python/sglang/srt/managers/multimodal_processor.py

@@ -12,8 +12,7 @@ logger = logging.getLogger(__name__)
 PROCESSOR_MAPPING = {}
 
 
-def import_processors():
-    package_name = "sglang.srt.multimodal.processors"
+def import_processors(package_name: str):
     package = importlib.import_module(package_name)
     for _, name, ispkg in pkgutil.iter_modules(package.__path__, package_name + "."):
         if not ispkg:

python/sglang/srt/managers/tokenizer_manager.py

@@ -185,7 +185,7 @@ class TokenizerManager(TokenizerCommunicatorMixin):
         )
 
         if self.model_config.is_multimodal:
-            import_processors()
+            import_processors("sglang.srt.multimodal.processors")
             try:
                 _processor = get_processor(
                     server_args.tokenizer_path,

python/sglang/srt/models/torch_native_llama.py

@@ -66,8 +66,8 @@ from sglang.srt.model_executor.forward_batch_info import ForwardBatch
 from sglang.srt.model_loader.weight_utils import default_weight_loader
 from sglang.srt.utils import add_prefix
 
-tp_size = get_tensor_model_parallel_world_size()
-tp_rank = get_tensor_model_parallel_rank()
+tp_size: Optional[int] = None
+tp_rank: Optional[int] = None
 
 
 def gate_up_proj_weight_loader(
@@ -341,6 +341,13 @@ class LlamaModel(nn.Module):
         quant_config: Optional[QuantizationConfig] = None,
     ) -> None:
         super().__init__()
+
+        global tp_size, tp_rank
+        if tp_size is None:
+            tp_size = get_tensor_model_parallel_world_size()
+        if tp_rank is None:
+            tp_rank = get_tensor_model_parallel_rank()
+
         self.config = config
         self.padding_idx = config.pad_token_id
         self.vocab_size = config.vocab_size

python/sglang/utils.py

@@ -8,6 +8,7 @@ import logging
 import os
 import random
 import socket
+import ssl
 import subprocess
 import sys
 import time
@@ -158,7 +159,15 @@ def http_request(
             data = bytes(dumps(json), encoding="utf-8")
 
         try:
-            resp = urllib.request.urlopen(req, data=data, cafile=verify)
+            if sys.version_info >= (3, 13):
+                # Python 3.13+: Use SSL context (cafile removed)
+                if verify and isinstance(verify, str):
+                    context = ssl.create_default_context(cafile=verify)
+                else:
+                    context = ssl.create_default_context()
+                resp = urllib.request.urlopen(req, data=data, context=context)
+            else:
+                resp = urllib.request.urlopen(req, data=data, cafile=verify)
             return HttpResponse(resp)
         except urllib.error.HTTPError as e:
             return HttpResponse(e)