[Test] split test/registered/distributed/ into topic folders (#26240)

2026-05-24 17:02:07 -07:00
parent 85471d253d
commit d7e3e54148
16 changed files with 1 additions and 1 deletions
@@ -0,0 +1,185 @@
+import unittest
+
+import requests
+
+from sglang.lang.chat_template import get_chat_template_by_model_path
+from sglang.srt.environ import envs
+from sglang.srt.utils import kill_process_tree
+from sglang.test.ci.ci_register import register_cuda_ci
+from sglang.test.kits.ebnf_constrained_kit import EBNFConstrainedMixin
+from sglang.test.kits.eval_accuracy_kit import GSM8KMixin
+from sglang.test.kits.json_constrained_kit import JSONConstrainedMixin
+from sglang.test.kits.radix_cache_server_kit import run_radix_attention_test
+from sglang.test.kits.regex_constrained_kit import RegexConstrainedMixin
+from sglang.test.test_utils import (
+    DEFAULT_IMAGE_URL,
+    DEFAULT_MLA_MODEL_NAME_FOR_TEST,
+    DEFAULT_MODEL_NAME_FOR_TEST_MLA,
+    DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+    DEFAULT_URL_FOR_TEST,
+    CustomTestCase,
+    popen_launch_server,
+)
+
+register_cuda_ci(est_time=420, stage="base-b", runner_config="2-gpu-large")
+
+
+class TestDPAttentionDP2TP2(
+    CustomTestCase,
+    GSM8KMixin,
+    JSONConstrainedMixin,
+    EBNFConstrainedMixin,
+    RegexConstrainedMixin,
+):
+    gsm8k_accuracy_thres = 0.6
+
+    @classmethod
+    def setUpClass(cls):
+        cls.model = DEFAULT_MODEL_NAME_FOR_TEST_MLA
+        cls.base_url = DEFAULT_URL_FOR_TEST
+        cls._env_override = envs.SGLANG_DISABLE_CONSECUTIVE_PREFILL_OVERLAP.override(
+            True
+        )
+        cls._env_override.__enter__()
+        cls.process = popen_launch_server(
+            cls.model,
+            cls.base_url,
+            timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+            other_args=[
+                "--trust-remote-code",
+                "--tp",
+                "2",
+                "--enable-dp-attention",
+                "--dp",
+                "2",
+                "--enable-torch-compile",
+                "--torch-compile-max-bs",
+                "2",
+            ],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+        cls._env_override.__exit__(None, None, None)
+
+
+class TestDPAttentionMixedChunk(
+    CustomTestCase,
+    GSM8KMixin,
+):
+    gsm8k_accuracy_thres = 0.6
+
+    @classmethod
+    def setUpClass(cls):
+        cls.model = DEFAULT_MLA_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",
+                "--tp",
+                "2",
+                "--enable-dp-attention",
+                "--dp",
+                "2",
+                "--enable-mixed-chunk",
+                "--chunked-prefill-size",
+                "256",
+            ],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+
+class TestDPRetract(
+    CustomTestCase,
+    JSONConstrainedMixin,
+    EBNFConstrainedMixin,
+    RegexConstrainedMixin,
+):
+    @classmethod
+    def setUpClass(cls):
+        cls.model = DEFAULT_MLA_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",
+                "--tp",
+                "2",
+                "--enable-dp-attention",
+                "--dp",
+                "2",
+                "--max-total-tokens",
+                "4500",
+                "--max-running-requests",
+                "128",
+                "--chunked-prefill-size",
+                "256",
+            ],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+    def test_radix_attention(self):
+        with envs.SGLANG_TEST_RETRACT.override(True):
+            run_radix_attention_test(self.base_url)
+            self.assertIsNone(self.process.poll())
+
+
+class TestDPAttentionDP2TP2VLM(CustomTestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.model = "moonshotai/Kimi-VL-A3B-Instruct"
+        cls.base_url = DEFAULT_URL_FOR_TEST
+        cls.image_url = DEFAULT_IMAGE_URL
+        cls.process = popen_launch_server(
+            cls.model,
+            cls.base_url,
+            timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
+            other_args=[
+                "--trust-remote-code",
+                "--tp",
+                "2",
+                "--enable-dp-attention",
+                "--dp",
+                "2",
+            ],
+        )
+
+    @classmethod
+    def tearDownClass(cls):
+        kill_process_tree(cls.process.pid)
+
+    def test_vlm_generate(self):
+        chat_template = get_chat_template_by_model_path(self.model)
+        prompt = f"{chat_template.image_token}What is in this image?"
+        response = requests.post(
+            self.base_url + "/generate",
+            json={
+                "text": prompt,
+                "image_data": [self.image_url],
+                "sampling_params": {
+                    "temperature": 0,
+                    "max_new_tokens": 16,
+                },
+            },
+        )
+        response.raise_for_status()
+        response_json = response.json()
+        print(response_json)
+        self.assertIn("output_ids", response_json)
+        self.assertGreater(len(response_json["output_ids"]), 0)
+
+
+if __name__ == "__main__":
+    unittest.main()