minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

[Test] Consolidate eval accuracy test mixins into eval_accuracy_kit (#21047)

Browse Source
This commit is contained in:
Liangsheng Yin
2026-03-26 14:26:46 -07:00
committed by GitHub
parent e5dd411f64
commit fb90c9d298
25 changed files with 276 additions and 377 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/mla/test_mla.py
+4 -15
View File
@@ -1,9 +1,8 @@
import unittest
from types import SimpleNamespace
from sglang.srt.utils import kill_process_tree
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
from sglang.test.run_eval import run_eval
from sglang.test.kits.eval_accuracy_kit import MGSMEnMixin
from sglang.test.test_utils import (
DEFAULT_MLA_MODEL_NAME_FOR_TEST,
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
@@ -17,7 +16,9 @@ register_cuda_ci(est_time=194, suite="stage-b-test-1-gpu-large")
register_amd_ci(est_time=1100, suite="stage-b-test-1-gpu-small-amd")
class TestMLA(CustomTestCase):
class TestMLA(CustomTestCase, MGSMEnMixin):
mgsm_en_score_threshold = 0.8
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MLA_MODEL_NAME_FOR_TEST
@@ -40,18 +41,6 @@ class TestMLA(CustomTestCase):
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def test_mgsm_en(self):
args = SimpleNamespace(
base_url=self.base_url,
model=self.model,
eval_name="mgsm_en",
num_examples=None,
num_threads=1024,
)
metrics = run_eval(args)
self.assertGreater(metrics["score"], 0.8)
if __name__ == "__main__":
unittest.main()
test/registered/mla/test_mla_fp8.py
+4 -15
View File
@@ -1,9 +1,8 @@
import unittest
from types import SimpleNamespace
from sglang.srt.utils import kill_process_tree
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
from sglang.test.run_eval import run_eval
from sglang.test.kits.eval_accuracy_kit import MGSMEnMixin
from sglang.test.test_utils import (
DEFAULT_MLA_FP8_MODEL_NAME_FOR_TEST,
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
@@ -17,7 +16,9 @@ register_cuda_ci(est_time=77, suite="stage-b-test-1-gpu-large")
register_amd_ci(est_time=800, suite="stage-b-test-1-gpu-small-amd")
class TestMLA(CustomTestCase):
class TestMLA(CustomTestCase, MGSMEnMixin):
mgsm_en_score_threshold = 0.8
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MLA_FP8_MODEL_NAME_FOR_TEST
@@ -37,18 +38,6 @@ class TestMLA(CustomTestCase):
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def test_mgsm_en(self):
args = SimpleNamespace(
base_url=self.base_url,
model=self.model,
eval_name="mgsm_en",
num_examples=None,
num_threads=1024,
)
metrics = run_eval(args)
assert metrics["score"] >= 0.8
if __name__ == "__main__":
unittest.main()