Add accuracy and latency tests of eagle into CI (#3027)

2025-01-21 02:55:14 -08:00
parent ec1c21cdc4
commit a4331cd260
7 changed files with 186 additions and 123 deletions
--- a/test/srt/test_torch_compile.py
+++ b/test/srt/test_torch_compile.py
@@ -23,7 +23,7 @@ class TestTorchCompile(unittest.TestCase):
            cls.model,
            cls.base_url,
            timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
-            other_args=["--enable-torch-compile"],
+            other_args=["--enable-torch-compile", "--cuda-graph-max-bs", "4"],
        )

    @classmethod