chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

Revert "tiny remove deprecated endpoint call" (#14533)

Browse Source
This commit is contained in:
Baizhou Zhang
2025-12-05 23:48:54 -08:00
committed by GitHub
parent d257bf87b9
commit 42fcf5438f
36 changed files with 59 additions and 59 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
test/manual/test_deepseek_v32_cp_single_node.py
View File

@@ -79,7 +79,7 @@ class TestDeepseekV32CP(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/manual/test_eagle_infer_beta_dp_attention.py
View File

@@ -77,7 +77,7 @@ class TestEagleDPAttnServerBase(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

20
test/manual/test_weight_version.py
View File

@@ -3,7 +3,7 @@ Test weight version functionality.
This test suite verifies the weight_version feature implementation including:
1. Default weight_version setting
2. /weight_version endpoint
2. /get_weight_version endpoint
3. /update_weight_version endpoint
4. /generate request meta_info contains weight_version
5. OpenAI API response metadata contains weight_version
@@ -48,13 +48,13 @@ class TestWeightVersion(CustomTestCase):
def test_weight_version_comprehensive(self):
"""Comprehensive test for all weight_version functionality."""
response = requests.get(f"{self.base_url}/model_info")
response = requests.get(f"{self.base_url}/get_model_info")
self.assertEqual(response.status_code, 200)
data = response.json()
self.assertIn("weight_version", data)
self.assertEqual(data["weight_version"], "test_version_1.0")
response = requests.get(f"{self.base_url}/weight_version")
response = requests.get(f"{self.base_url}/get_weight_version")
self.assertEqual(response.status_code, 200)
data = response.json()
self.assertIn("weight_version", data)
@@ -114,7 +114,7 @@ class TestWeightVersion(CustomTestCase):
self.assertTrue(data["success"])
self.assertEqual(data["new_version"], "updated_version_2.0")
response = requests.get(f"{self.base_url}/weight_version")
response = requests.get(f"{self.base_url}/get_weight_version")
self.assertEqual(response.status_code, 200)
data = response.json()
self.assertEqual(data["weight_version"], "updated_version_2.0")
@@ -148,13 +148,13 @@ class TestWeightVersion(CustomTestCase):
self.assertTrue(data["success"])
self.assertEqual(data["new_version"], "final_version_3.0")
# Check /weight_version
response = requests.get(f"{self.base_url}/weight_version")
# Check /get_weight_version
response = requests.get(f"{self.base_url}/get_weight_version")
self.assertEqual(response.status_code, 200)
self.assertEqual(response.json()["weight_version"], "final_version_3.0")
# Check /model_info
response = requests.get(f"{self.base_url}/model_info")
# Check /get_model_info
response = requests.get(f"{self.base_url}/get_model_info")
self.assertEqual(response.status_code, 200)
self.assertEqual(response.json()["weight_version"], "final_version_3.0")
@@ -193,7 +193,7 @@ class TestWeightVersion(CustomTestCase):
print("Testing weight_version update with real weight operations...")
# Get current model info for reference
model_info_response = requests.get(f"{self.base_url}/model_info")
model_info_response = requests.get(f"{self.base_url}/get_model_info")
self.assertEqual(model_info_response.status_code, 200)
current_model_path = model_info_response.json()["model_path"]
@@ -214,7 +214,7 @@ class TestWeightVersion(CustomTestCase):
)
# Verify version was updated
version_response = requests.get(f"{self.base_url}/weight_version")
version_response = requests.get(f"{self.base_url}/get_weight_version")
self.assertEqual(version_response.status_code, 200)
self.assertEqual(
version_response.json()["weight_version"], "disk_update_v2.0.0"

2
test/srt/ep/test_deepep_large.py
View File

@@ -143,7 +143,7 @@ class TestDeepseekMTP(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.92)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

4
test/srt/ep/test_deepep_small.py
View File

@@ -307,7 +307,7 @@ class TestMTP(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]
@@ -381,7 +381,7 @@ class TestMTPWithTBO(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/hicache/test_hicache_variants.py
View File

@@ -152,7 +152,7 @@ class TestHiCacheEagle(HiCacheBaseServer, HiCacheEvalMixin):
self.assertGreaterEqual(metrics["score"], self.expected_mmlu_score)
# EAGLE-specific check
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
print(f"{server_info=}")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"

2
test/srt/quant/test_w4a8_deepseek_v3.py
View File

@@ -103,7 +103,7 @@ class TestDeepseekV3W4Afp8Mtp(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

6
test/srt/rl/test_update_weights_from_disk.py
View File

@@ -113,7 +113,7 @@ class TestServerUpdateWeightsFromDisk(CustomTestCase):
return response.json()
def get_model_info(self):
response = requests.get(self.base_url + "/model_info")
response = requests.get(self.base_url + "/get_model_info")
model_path = response.json()["model_path"]
print(json.dumps(response.json()))
return model_path
@@ -254,7 +254,7 @@ class TestServerUpdateWeightsFromDiskAbortAllRequests(CustomTestCase):
return response.json()
def get_model_info(self):
response = requests.get(self.base_url + "/model_info")
response = requests.get(self.base_url + "/get_model_info")
model_path = response.json()["model_path"]
print(json.dumps(response.json()))
return model_path
@@ -386,7 +386,7 @@ class TestUpdateWeightsFromDiskParameterized(CustomTestCase):
return response.json()["text"]
def get_model_info():
response = requests.get(base_url + "/model_info")
response = requests.get(base_url + "/get_model_info")
model_path = response.json()["model_path"]
print(json.dumps(response.json()))
return model_path

2
test/srt/rl/test_update_weights_from_tensor.py
View File

@@ -209,7 +209,7 @@ class TestServerUpdateWeightsFromTensorNonBlocking(CustomTestCase):
return response.json()
def get_model_info(self):
response = requests.get(self.base_url + "/model_info")
response = requests.get(self.base_url + "/get_model_info")
model_path = response.json()["model_path"]
print(json.dumps(response.json()))
return model_path

4
test/srt/test_data_parallelism.py
View File

@@ -65,12 +65,12 @@ class TestDataParallelism(CustomTestCase):
def test_get_memory_pool_size(self):
# use `get_server_info` instead since `get_memory_pool_size` is merged into `get_server_info`
response = requests.get(self.base_url + "/server_info")
response = requests.get(self.base_url + "/get_server_info")
assert response.status_code == 200
time.sleep(1)
response = requests.get(self.base_url + "/server_info")
response = requests.get(self.base_url + "/get_server_info")
assert response.status_code == 200

2
test/srt/test_deepseek_v32_mtp.py
View File

@@ -69,7 +69,7 @@ class TestDeepseekV32MTP(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_deepseek_v3_fp4_4gpu.py
View File

@@ -143,7 +143,7 @@ class TestDeepseekV3FP4MTP(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_deepseek_v3_mtp.py
View File

@@ -67,7 +67,7 @@ class TestDeepseekV3MTP(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_dp_attention.py
View File

@@ -112,7 +112,7 @@ class TestDPAttentionDP2TP2DeepseekV3MTP(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_eagle_dp_attention.py
View File

@@ -78,7 +78,7 @@ class TestEAGLE3EngineDPAttention(CustomTestCase):
metrics = run_eval_few_shot_gsm8k(args)
print(f"{metrics=}")
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
server_data = server_info.json()
# Try to get avg_spec_accept_length

2
test/srt/test_eagle_infer_b.py
View File

@@ -144,7 +144,7 @@ class TestEAGLEServer(CustomTestCase):
print(f"{metrics=}")
self.assertGreater(metrics["accuracy"], 0.20)
server_info = requests.get(self.base_url + "/server_info").json()
server_info = requests.get(self.base_url + "/get_server_info").json()
avg_spec_accept_length = server_info["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_fa3.py
View File

@@ -112,7 +112,7 @@ class BaseFlashAttentionTest(CustomTestCase):
self.assertGreater(metrics[metric_key], self.accuracy_threshold)
if self.speculative_decode:
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_flashmla.py
View File

@@ -119,7 +119,7 @@ class TestFlashMLAMTP(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
print(f"{server_info=}")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"

2
test/srt/test_hybrid_attn_backend.py
View File

@@ -88,7 +88,7 @@ class TestHybridAttnBackendBase(CustomTestCase):
self.assertGreater(metrics[metric_key], self.accuracy_threshold)
if self.speculative_decode:
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_mla_deepseek_v3.py
View File

@@ -179,7 +179,7 @@ class TestDeepseekV3MTP(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_mla_flashinfer.py
View File

@@ -111,7 +111,7 @@ class TestFlashinferMLAMTP(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
print(f"{server_info=}")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"

4
test/srt/test_mla_int8_deepseek_v3.py
View File

@@ -113,7 +113,7 @@ class TestDeepseekV3MTPChannelInt8(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]
@@ -217,7 +217,7 @@ class TestDeepseekV3MTPBlockInt8(CustomTestCase):
self.assertGreater(metrics["accuracy"], 0.60)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

2
test/srt/test_ngram_speculative_decoding.py
View File

@@ -80,7 +80,7 @@ class TestNgramSpeculativeDecodingBase(CustomTestCase):
metric_key = "accuracy"
self.assertGreater(metrics[metric_key], self.accuracy_threshold)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]

4
test/srt/test_srt_endpoint.py
View File

@@ -496,7 +496,7 @@ class TestSRTEndpoint(CustomTestCase):
self.assertEqual(send_and_check_cached_tokens(range(0, 11000)), 10000)
def test_get_server_info(self):
response = requests.get(self.base_url + "/server_info")
response = requests.get(self.base_url + "/get_server_info")
response_json = response.json()
max_total_num_tokens = response_json["max_total_num_tokens"]
@@ -626,7 +626,7 @@ class TestSRTEndpoint(CustomTestCase):
tp = ThreadPoolExecutor(max_workers=30)
def s():
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
server_info.json()
futures = []

2
test/srt/test_standalone_speculative_decoding.py
View File

@@ -88,7 +88,7 @@ class TestStandaloneSpeculativeDecodingBase(CustomTestCase):
metric_key = "accuracy"
self.assertGreater(metrics[metric_key], self.accuracy_threshold)
server_info = requests.get(self.base_url + "/server_info")
server_info = requests.get(self.base_url + "/get_server_info")
avg_spec_accept_length = server_info.json()["internal_states"][0][
"avg_spec_accept_length"
]