66 lines
1.8 KiB
Python
66 lines
1.8 KiB
Python
import os
|
|
import unittest
|
|
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.test_utils import (
|
|
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
|
|
DEFAULT_URL_FOR_TEST,
|
|
CustomTestCase,
|
|
popen_launch_server,
|
|
)
|
|
|
|
register_cuda_ci(est_time=300, suite="nightly-4-gpu-b200", nightly=True)
|
|
|
|
|
|
class TestFlashinferTrtllmGenAttnBackend(CustomTestCase):
|
|
@classmethod
|
|
def setUpClass(cls):
|
|
cls.model = "Qwen/Qwen3-Next-80B-A3B-Instruct"
|
|
cls.base_url = DEFAULT_URL_FOR_TEST
|
|
cls.process = popen_launch_server(
|
|
cls.model,
|
|
cls.base_url,
|
|
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
|
|
env={**os.environ, "SGLANG_ENABLE_JIT_DEEPGEMM": "False"},
|
|
other_args=[
|
|
"--attention-backend",
|
|
"trtllm_mha",
|
|
"--cuda-graph-max-bs-decode",
|
|
"512",
|
|
"--tp-size",
|
|
"4",
|
|
"--ep-size",
|
|
"4",
|
|
"--mem-fraction-static",
|
|
"0.7",
|
|
"--mamba-ssm-dtype",
|
|
"bfloat16",
|
|
"--disable-radix-cache",
|
|
],
|
|
)
|
|
|
|
@classmethod
|
|
def tearDownClass(cls):
|
|
kill_process_tree(cls.process.pid)
|
|
|
|
def test_gsm8k(self):
|
|
args = SimpleNamespace(
|
|
base_url=self.base_url,
|
|
model=self.model,
|
|
eval_name="gsm8k",
|
|
api="completion",
|
|
max_tokens=512,
|
|
num_examples=200,
|
|
num_threads=128,
|
|
)
|
|
metrics = run_eval(args)
|
|
print(f"{metrics=}")
|
|
self.assertGreater(metrics["score"], 0.93)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|