apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[Metax][CI] e2e ci tests enable cuda graph (#6401)

Browse Source
This commit is contained in:
MingkunZhang
2026-02-09 16:25:23 +08:00
committed by GitHub
parent fd56d85346
commit 268276e287
4 changed files with 11 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/metax_ci/test_ernie_21b.py
+1 -1
View File
@@ -31,7 +31,7 @@ class TestErnie21B(unittest.TestCase):
load_choices="default_v1",
# enable_prefix_caching=False,
disable_custom_all_reduce=True,
graph_optimization_config={"use_cudagraph": False, "graph_opt_level": 0},
# graph_optimization_config={"use_cudagraph": False, "graph_opt_level": 0},
)
cls.sampling_params = fastdeploy.SamplingParams(top_p=0.95, max_tokens=256, temperature=0.6)
tests/metax_ci/test_ernie_28b_vl.py
+1 -1
View File
@@ -37,7 +37,7 @@ class TestErnie28BVL(unittest.TestCase):
quantization="wint8",
disable_custom_all_reduce=True,
# enable_prefix_caching=False,
graph_optimization_config={"use_cudagraph": False, "graph_opt_level": 0},
# graph_optimization_config={"use_cudagraph": False, "graph_opt_level": 0},
limit_mm_per_prompt={"image": 100},
reasoning_parser="ernie-45-vl",
load_choices="default_v1",