Files
sglang/test/registered/backends/test_torch_compile.py

85 lines
2.3 KiB
Python

import time
import unittest
from types import SimpleNamespace
import requests
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_MODEL_NAME_FOR_TEST,
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
DEFAULT_URL_FOR_TEST,
CustomTestCase,
is_in_amd_ci,
popen_launch_server,
)
register_cuda_ci(est_time=144, suite="stage-b-test-large-1-gpu")
register_amd_ci(est_time=1100, suite="stage-b-test-small-1-gpu-amd")
class TestTorchCompile(CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_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=["--enable-torch-compile", "--cuda-graph-max-bs", "4"],
)
@classmethod
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def test_mmlu(self):
args = SimpleNamespace(
base_url=self.base_url,
model=self.model,
eval_name="mmlu",
num_examples=64,
num_threads=32,
)
metrics = run_eval(args)
self.assertGreaterEqual(metrics["score"], 0.65)
def run_decode(self, max_new_tokens):
response = requests.post(
self.base_url + "/generate",
json={
"text": "The capital of France is",
"sampling_params": {
"temperature": 0,
"max_new_tokens": max_new_tokens,
"ignore_eos": True,
},
},
)
return response.json()
def test_throughput(self):
# Warmup
res = self.run_decode(16)
max_tokens = 256
tic = time.perf_counter()
res = self.run_decode(max_tokens)
tok = time.perf_counter()
print(f"{res=}")
throughput = max_tokens / (tok - tic)
print(f"Throughput: {throughput} tokens/s")
if is_in_amd_ci():
self.assertGreaterEqual(throughput, 145)
else:
self.assertGreaterEqual(throughput, 152)
if __name__ == "__main__":
unittest.main()