[test] Add GSM8K accuracy test for PP with mixed chunk prefill (#23029)
Co-authored-by: cyy <cy02433585@alibaba-inc.com>
This commit is contained in:
@@ -4,6 +4,7 @@ python3 -m unittest test_pp_single_node.TestPPAccuracy.test_gsm8k
|
|||||||
python3 -m unittest test_pp_single_node.TestQwenPPAccuracy.test_pp_consistency
|
python3 -m unittest test_pp_single_node.TestQwenPPAccuracy.test_pp_consistency
|
||||||
python3 -m unittest test_pp_single_node.TestFixedBugs.test_chunked_prefill_with_small_bs
|
python3 -m unittest test_pp_single_node.TestFixedBugs.test_chunked_prefill_with_small_bs
|
||||||
python3 -m unittest test_pp_single_node.TestQwenVLPPAccuracy.test_mmmu
|
python3 -m unittest test_pp_single_node.TestQwenVLPPAccuracy.test_mmmu
|
||||||
|
python3 -m unittest test_pp_single_node.TestPPMixedChunk.test_gsm8k
|
||||||
"""
|
"""
|
||||||
|
|
||||||
import time
|
import time
|
||||||
@@ -420,6 +421,53 @@ class TestQwen35PPAccuracy(unittest.TestCase):
|
|||||||
)
|
)
|
||||||
|
|
||||||
|
|
||||||
|
class TestPPMixedChunk(CustomTestCase):
|
||||||
|
@classmethod
|
||||||
|
def setUpClass(cls):
|
||||||
|
cls.model = DEFAULT_MODEL_NAME_FOR_TEST
|
||||||
|
cls.base_url = "http://127.0.0.1:23338"
|
||||||
|
cls.process = popen_launch_server(
|
||||||
|
cls.model,
|
||||||
|
cls.base_url,
|
||||||
|
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
|
||||||
|
other_args=[
|
||||||
|
"--tp-size",
|
||||||
|
2,
|
||||||
|
"--pp-size",
|
||||||
|
2,
|
||||||
|
"--chunked-prefill-size",
|
||||||
|
256,
|
||||||
|
"--enable-mixed-chunk",
|
||||||
|
],
|
||||||
|
)
|
||||||
|
|
||||||
|
@classmethod
|
||||||
|
def tearDownClass(cls):
|
||||||
|
if hasattr(cls, "process"):
|
||||||
|
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=}")
|
||||||
|
|
||||||
|
if is_in_amd_ci():
|
||||||
|
# AMD triton backend produces slightly lower accuracy than FA3 on NVIDIA
|
||||||
|
self.assertGreater(metrics["score"], 0.70)
|
||||||
|
else:
|
||||||
|
self.assertGreater(metrics["score"], 0.74)
|
||||||
|
# Wait a little bit so that the memory check happens.
|
||||||
|
time.sleep(4)
|
||||||
|
|
||||||
|
|
||||||
class TestFixedBugs(unittest.TestCase):
|
class TestFixedBugs(unittest.TestCase):
|
||||||
def test_chunked_prefill_with_small_bs(self):
|
def test_chunked_prefill_with_small_bs(self):
|
||||||
model = DEFAULT_MODEL_NAME_FOR_TEST
|
model = DEFAULT_MODEL_NAME_FOR_TEST
|
||||||
@@ -435,7 +483,7 @@ class TestFixedBugs(unittest.TestCase):
|
|||||||
2,
|
2,
|
||||||
"--pp-size",
|
"--pp-size",
|
||||||
2,
|
2,
|
||||||
"--chunked-prefill",
|
"--chunked-prefill-size",
|
||||||
256,
|
256,
|
||||||
"--max-running-requests",
|
"--max-running-requests",
|
||||||
2,
|
2,
|
||||||
|
|||||||
Reference in New Issue
Block a user