minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[diffusion] feat: add performance mode server args (#24491)

Browse Source
This commit is contained in:
Mick
2026-05-14 00:57:46 +08:00
committed by GitHub
parent e2290b155a
commit ff70aeac30
79 changed files with 1734 additions and 532 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/manual/test_tokenizer_manager.py
+28 -20
View File
@@ -37,11 +37,13 @@ class TestInputFormatDetection(unittest.TestCase):
self.server_args = ServerArgs(model_path=DEFAULT_SMALL_MODEL_NAME_FOR_TEST)
self.port_args = PortArgs.init_new(self.server_args)
with patch("zmq.asyncio.Context"), patch(
"sglang.srt.utils.network.get_zmq_socket"
), patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer:
with (
patch("zmq.asyncio.Context"),
patch("sglang.srt.utils.network.get_zmq_socket"),
patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer,
):
mock_tokenizer.return_value = Mock(vocab_size=32000)
self.tokenizer_manager = TokenizerManager(self.server_args, self.port_args)
@@ -133,11 +135,13 @@ class TestTokenizerInputPreparation(unittest.TestCase):
self.server_args = ServerArgs(model_path=DEFAULT_SMALL_MODEL_NAME_FOR_TEST)
self.port_args = PortArgs.init_new(self.server_args)
with patch("zmq.asyncio.Context"), patch(
"sglang.srt.utils.network.get_zmq_socket"
), patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer:
with (
patch("zmq.asyncio.Context"),
patch("sglang.srt.utils.network.get_zmq_socket"),
patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer,
):
mock_tokenizer.return_value = Mock(vocab_size=32000)
self.tokenizer_manager = TokenizerManager(self.server_args, self.port_args)
@@ -191,11 +195,13 @@ class TestTokenizerResultExtraction(unittest.TestCase):
self.server_args = ServerArgs(model_path=DEFAULT_SMALL_MODEL_NAME_FOR_TEST)
self.port_args = PortArgs.init_new(self.server_args)
with patch("zmq.asyncio.Context"), patch(
"sglang.srt.utils.network.get_zmq_socket"
), patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer:
with (
patch("zmq.asyncio.Context"),
patch("sglang.srt.utils.network.get_zmq_socket"),
patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer,
):
mock_tokenizer.return_value = Mock(vocab_size=32000)
self.tokenizer_manager = TokenizerManager(self.server_args, self.port_args)
@@ -313,11 +319,13 @@ class TestTokenizerManagerIntegration(unittest.TestCase):
self.server_args = ServerArgs(model_path=DEFAULT_SMALL_MODEL_NAME_FOR_TEST)
self.port_args = PortArgs.init_new(self.server_args)
with patch("zmq.asyncio.Context"), patch(
"sglang.srt.utils.network.get_zmq_socket"
), patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer:
with (
patch("zmq.asyncio.Context"),
patch("sglang.srt.utils.network.get_zmq_socket"),
patch(
"sglang.srt.utils.hf_transformers_utils.get_tokenizer"
) as mock_tokenizer,
):
mock_tokenizer.return_value = Mock(vocab_size=32000)
self.tokenizer_manager = TokenizerManager(self.server_args, self.port_args)