[CI]Remove msgm-en and mmlu tests which cause timeout (#21733)

This commit is contained in:
Baizhou Zhang
2026-03-31 01:10:05 -07:00
committed by GitHub
parent 5628e908ae
commit d52757fe97
6 changed files with 43 additions and 42 deletions
+26 -20
View File
@@ -3,9 +3,9 @@ from types import SimpleNamespace
from sglang.srt.utils import kill_process_tree
from sglang.test.ci.ci_register import register_cuda_ci
from sglang.test.run_eval import run_eval
from sglang.test.few_shot_gsm8k import run_eval as run_eval_few_shot_gsm8k
from sglang.test.test_utils import (
DEFAULT_MLA_MODEL_NAME_FOR_TEST,
DEFAULT_MODEL_NAME_FOR_TEST_MLA,
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
DEFAULT_URL_FOR_TEST,
CustomTestCase,
@@ -18,7 +18,7 @@ register_cuda_ci(est_time=140, suite="stage-b-test-2-gpu-large")
class TestEp(CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MLA_MODEL_NAME_FOR_TEST
cls.model = DEFAULT_MODEL_NAME_FOR_TEST_MLA
cls.base_url = DEFAULT_URL_FOR_TEST
cls.process = popen_launch_server(
cls.model,
@@ -37,23 +37,26 @@ class TestEp(CustomTestCase):
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def test_mgsm_en(self):
def test_gsm8k(self):
args = SimpleNamespace(
base_url=self.base_url,
model=self.model,
eval_name="mgsm_en",
num_examples=None,
num_threads=1024,
num_shots=5,
data_path=None,
num_questions=200,
max_new_tokens=512,
parallel=128,
host="http://127.0.0.1",
port=int(self.base_url.split(":")[-1]),
)
metrics = run_eval_few_shot_gsm8k(args)
print(metrics)
metrics = run_eval(args)
self.assertGreaterEqual(metrics["score"], 0.8)
self.assertGreater(metrics["accuracy"], 0.60)
class TestEpDeepGEMM(CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MLA_MODEL_NAME_FOR_TEST
cls.model = DEFAULT_MODEL_NAME_FOR_TEST_MLA
cls.base_url = DEFAULT_URL_FOR_TEST
cls.process = popen_launch_server(
cls.model,
@@ -76,17 +79,20 @@ class TestEpDeepGEMM(CustomTestCase):
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def test_mgsm_en(self):
def test_gsm8k(self):
args = SimpleNamespace(
base_url=self.base_url,
model=self.model,
eval_name="mgsm_en",
num_examples=None,
num_threads=1024,
num_shots=5,
data_path=None,
num_questions=200,
max_new_tokens=512,
parallel=128,
host="http://127.0.0.1",
port=int(self.base_url.split(":")[-1]),
)
metrics = run_eval_few_shot_gsm8k(args)
print(metrics)
metrics = run_eval(args)
self.assertGreaterEqual(metrics["score"], 0.8)
self.assertGreater(metrics["accuracy"], 0.60)
if __name__ == "__main__":