ci: prune per-commit CUDA tests — move 25 files + 13 testcases to test/manual/ (#24721)

2026-05-08 15:53:23 -07:00
parent aefd8e257f
commit 5fbec0e445
45 changed files with 818 additions and 525 deletions
@@ -0,0 +1,83 @@
+# Qwen model tests
+
+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.test_utils import (
+    DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+    DEFAULT_URL_FOR_TEST,
+    CustomTestCase,
+    popen_launch_server,
+)
+
+register_cuda_ci(est_time=108, suite="stage-b-test-1-gpu-small")
+register_amd_ci(est_time=130, suite="stage-b-test-1-gpu-small-amd")
+
+
+class TestQwen2(CustomTestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.model = "Qwen/Qwen2-7B-Instruct"
+        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=[],
+        )
+
+    @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"], 0.78)
+
+
+class TestQwen2FP8(CustomTestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.model = "neuralmagic/Qwen2-7B-Instruct-FP8"
+        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=[],
+        )
+
+    @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"], 0.78)
+
+
+if __name__ == "__main__":
+    unittest.main()