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

Use smaller embedding model when not testing model specifically (#13891)

Browse Source
This commit is contained in:
Harry Mellor
2025-02-28 08:50:43 +00:00
committed by GitHub
parent b9e41734c5
commit 76c89fcadd
9 changed files with 15 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/model_executor/test_model_load_with_params.py
View File

@@ -14,7 +14,7 @@ MODEL_NAME = os.environ.get("MODEL_NAME", "BAAI/bge-base-en-v1.5")
REVISION = os.environ.get("REVISION", "main")
MODEL_NAME_ROBERTA = os.environ.get("MODEL_NAME",
"intfloat/multilingual-e5-large")
"intfloat/multilingual-e5-small")
REVISION_ROBERTA = os.environ.get("REVISION", "main")
@@ -83,7 +83,7 @@ def test_roberta_model_loading_with_params(vllm_runner):
assert model_config.pooler_config.pooling_norm
# asserts on the tokenizer loaded
assert model_tokenizer.tokenizer_id == "intfloat/multilingual-e5-large"
assert model_tokenizer.tokenizer_id == "intfloat/multilingual-e5-small"
assert not model_tokenizer.tokenizer_config["do_lower_case"]
def check_model(model):