Unify GSM8K eval path to Chat API for regression CI readiness (#21667)

test/manual/ep/test_moe_deepep_eval_accuracy_large.py

@@ -7,7 +7,6 @@ import unittest
 from types import SimpleNamespace
 
 from sglang.srt.utils import kill_process_tree
-from sglang.test.few_shot_gsm8k import run_eval as run_eval_few_shot_gsm8k
 from sglang.test.run_eval import run_eval
 from sglang.test.test_utils import (
     DEFAULT_DEEPEP_MODEL_NAME_FOR_TEST,
@@ -44,18 +43,19 @@ class TestMoEDeepEPEvalAccuracyLarge(CustomTestCase):
 
     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=64,
             num_shots=8,
-            data_path=None,
-            num_questions=200,
-            parallel=64,
-            max_new_tokens=512,
-            host="http://127.0.0.1",
-            port=int(self.base_url.split(":")[-1]),
         )
-        metrics = run_eval_few_shot_gsm8k(args)
+        metrics = run_eval(args)
         print(f"Eval accuracy of GSM8K: {metrics=}")
 
-        self.assertGreater(metrics["accuracy"], 0.93)
+        self.assertGreater(metrics["score"], 0.93)
 
     def test_mmlu(self):
         args = SimpleNamespace(