c781db0f6c
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: cy <chenyang08056032@163.com>
34 lines
978 B
Python
34 lines
978 B
Python
import unittest
|
|
|
|
from sglang.test.ascend.gsm8k_ascend_mixin import GSM8KAscendMixin
|
|
from sglang.test.ascend.test_ascend_utils import EXAONE_3_5_7_8B_INSTRUCT_WEIGHTS_PATH
|
|
from sglang.test.ci.ci_register import register_npu_ci
|
|
from sglang.test.test_utils import CustomTestCase
|
|
|
|
register_npu_ci(est_time=400, suite="nightly-1-npu-a3", nightly=True)
|
|
|
|
|
|
class TestEXAONE(GSM8KAscendMixin, CustomTestCase):
|
|
"""Testcase: Verify that the inference accuracy of the LGAI-EXAONE/EXAONE-3.5-7.8B-Instruct model on the GSM8K dataset is no less than 0.8.
|
|
|
|
[Test Category] Model
|
|
[Test Target] LGAI-EXAONE/EXAONE-3.5-7.8B-Instruct
|
|
"""
|
|
|
|
model = EXAONE_3_5_7_8B_INSTRUCT_WEIGHTS_PATH
|
|
accuracy = 0.8
|
|
other_args = [
|
|
"--trust-remote-code",
|
|
"--mem-fraction-static",
|
|
"0.8",
|
|
"--attention-backend",
|
|
"ascend",
|
|
"--disable-cuda-graph",
|
|
"--dtype",
|
|
"bfloat16",
|
|
]
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|