biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Support embedding models in V1 (#16188)

Browse Source 
Signed-off-by: Max de Bayser <mbayser@br.ibm.com>
Signed-off-by: Max de Bayser <maxdebayser@gmail.com>
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
Co-authored-by: 22quinn <33176974+22quinn@users.noreply.github.com>
This commit is contained in:
Maximilien de Bayser
2025-06-19 01:36:33 -03:00
committed by GitHub
parent 4959915089
commit 799397ee4f
56 changed files with 889 additions and 281 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
tests/models/language/pooling/test_classification.py
View File

@@ -6,6 +6,14 @@ from transformers import AutoModelForSequenceClassification
from vllm.platforms import current_platform
# TODO: enable when float32 is supported by V1
# @pytest.fixture(autouse=True)
# def v1(run_with_both_engines):
# # Simple autouse wrapper to run both engines for each test
# # This can be promoted up to conftest.py to run for every
# # test in a package
# pass
@pytest.mark.parametrize(
"model",
@@ -29,7 +37,7 @@ def test_models(
# switch to use ROCm CK FA backend
monkeypatch.setenv("VLLM_USE_TRITON_FLASH_ATTN", "False")
with vllm_runner(model, dtype=dtype) as vllm_model:
with vllm_runner(model, max_model_len=512, dtype=dtype) as vllm_model:
vllm_outputs = vllm_model.classify(example_prompts)
with hf_runner(model,