import unittest from types import SimpleNamespace from sglang.srt.utils import kill_process_tree from sglang.test.run_eval import run_eval from sglang.test.test_utils import ( DEFAULT_MLA_FP8_MODEL_NAME_FOR_TEST, DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH, DEFAULT_URL_FOR_TEST, CustomTestCase, popen_launch_server, ) class TestMLA(CustomTestCase): @classmethod def setUpClass(cls): cls.model = DEFAULT_MLA_FP8_MODEL_NAME_FOR_TEST cls.base_url = DEFAULT_URL_FOR_TEST cls.process = popen_launch_server( cls.model, cls.base_url, timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH, other_args=[ "--trust-remote-code", "--kv-cache-dtype", "fp8_e5m2", ], ) @classmethod 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()