Migrate performance, accuracy, and quantization tests to CI registry (#17177)

Co-authored-by: Kangyan-Zhou <zky314343421@gmail.com>

test/registered/quant/test_awq.py

@@ -0,0 +1,114 @@
+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_AWQ_MOE_MODEL_NAME_FOR_TEST,
+    DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+    DEFAULT_URL_FOR_TEST,
+    CustomTestCase,
+    popen_launch_server,
+)
+
+register_cuda_ci(est_time=163, suite="stage-b-test-large-1-gpu")
+
+
+class TestAWQ(CustomTestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.model = DEFAULT_AWQ_MOE_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"],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+    def test_mmlu(self):
+        args = SimpleNamespace(
+            base_url=self.base_url,
+            model=self.model,
+            eval_name="mmlu",
+            num_examples=64,
+            num_threads=32,
+        )
+
+        metrics = run_eval(args)
+        self.assertGreater(metrics["score"], 0.64)
+
+
+class TestAWQMarlinBfloat16(CustomTestCase):
+    """
+    Verify that the model can be loaded with bfloat16 dtype and awq_marlin quantization
+    """
+
+    @classmethod
+    def setUpClass(cls):
+        cls.model = "QuantTrio/Qwen3-VL-30B-A3B-Instruct-AWQ"
+        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=["--dtype", "bfloat16", "--quantization", "awq_marlin"],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+    def test_mmlu(self):
+        args = SimpleNamespace(
+            base_url=self.base_url,
+            model=self.model,
+            eval_name="mmlu",
+            num_examples=64,
+            num_threads=32,
+        )
+
+        metrics = run_eval(args)
+        self.assertGreater(metrics["score"], 0.83)
+
+
+class TestAWQMarlinFloat16(CustomTestCase):
+    """
+    Verify that the model can be loaded with float16 dtype and awq_marlin quantization
+    """
+
+    @classmethod
+    def setUpClass(cls):
+        cls.model = "QuantTrio/Qwen3-VL-30B-A3B-Instruct-AWQ"
+        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=["--dtype", "float16", "--quantization", "awq_marlin"],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+    def test_mmlu(self):
+        args = SimpleNamespace(
+            base_url=self.base_url,
+            model=self.model,
+            eval_name="mmlu",
+            num_examples=64,
+            num_threads=32,
+        )
+
+        metrics = run_eval(args)
+        self.assertGreater(metrics["score"], 0.85)
+
+
+if __name__ == "__main__":
+    unittest.main()