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

[FEAT] [ROCm] [Embedding] Add encoder-only model support into ROCm Flash Attention to enable embedding models. (#14664)

Browse Source 
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
This commit is contained in:
TJian
2025-03-13 00:31:19 +08:00
committed by GitHub
parent d9f83d6206
commit 916836bbfb
7 changed files with 118 additions and 50 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/models/embedding/language/test_gritlm.py
View File

@@ -15,8 +15,8 @@ import vllm.config
from ....utils import RemoteOpenAIServer
# GritLM embedding implementation is only supported by XFormers backend.
pytest.mark.skipif(not importlib.util.find_spec("xformers"),
reason="GritLM requires XFormers")
pytestmark = pytest.mark.skipif(not importlib.util.find_spec("xformers"),
reason="GritLM requires XFormers")
MODEL_NAME = "parasail-ai/GritLM-7B-vllm"
MAX_MODEL_LEN = 4000