xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-09 22:44:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CI] Enable encoder model compilation test (#24442)

Browse Source 
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
This commit is contained in:
Jiangyun Zhu 2025-09-09 02:48:06 +08:00 committed by GitHub
parent 8d7f39b48c
commit 7be141b2c5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 17 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
tests/compile/test_basic_correctness.py
View File

@ -62,8 +62,12 @@ class TestSetting:
TestSetting(
model="BAAI/bge-multilingual-gemma2",
model_args=[
"--runner", "pooling", "--dtype", "bfloat16",
"--max-model-len", "2048"
"--runner",
"pooling",
"--dtype",
"bfloat16",
"--max-model-len",
"2048",
],
pp_size=1,
tp_size=1,
@ -71,17 +75,15 @@ class TestSetting:
method="encode",
fullgraph=True,
),
# TODO: bert models are not supported in V1 yet
# # encoder-based embedding model (BERT)
# TestSetting(
# model="BAAI/bge-base-en-v1.5",
# model_args=["--runner", "pooling"],
# pp_size=1,
# tp_size=1,
# attn_backend="XFORMERS",
# method="encode",
# fullgraph=True,
# ),
TestSetting(
model="BAAI/bge-base-en-v1.5",
model_args=["--runner", "pooling"],
pp_size=1,
tp_size=1,
attn_backend="FLASH_ATTN",
method="encode",
fullgraph=True,
),
# vision language model
TestSetting(
model="microsoft/Phi-3.5-vision-instruct",
@ -92,7 +94,8 @@ class TestSetting:
method="generate_with_image",
fullgraph=False,
),
])
],
)
def test_compile_correctness(
monkeypatch: pytest.MonkeyPatch,
test_setting: TestSetting,