Skip to content

fix: reject case-insensitive duplicate cross-encoder registrations #795

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
joein merged 2 commits into from
Oct 12, 2026
Merged

fix: reject case-insensitive duplicate cross-encoder registrations #795

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2a859a5
fix: 阻止重排模型大小写重复注册
wToTw123 Oct 9, 2026
a8dfe94
tests: move cross-encoder case-duplicate tests into test_custom_models
joein Oct 12, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion fastembed/rerank/cross_encoder/text_cross_encoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,9 +143,10 @@ def add_custom_model(
size_in_gb: float = 0.0,
additional_files: list[str] | None = None,
) -> None:
"""Register a custom reranker under a case-insensitively unique model name."""
registered_models = cls._list_supported_models()
for registered_model in registered_models:
if model == registered_model.model:
if model.lower() == registered_model.model.lower():
raise ValueError(
f"Model {model} is already registered in CrossEncoderModel, if you still want to add this model, "
f"please use another model name"
Expand Down
43 changes: 31 additions & 12 deletions tests/test_custom_models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -284,6 +284,21 @@ def test_custom_text_model_lookup_is_case_insensitive():
assert model.model._normalization is True


def test_custom_cross_encoder_lookup_is_case_insensitive():
model_name = "Org/Model"

TextCrossEncoder.add_custom_model(
model_name,
sources=ModelSource(hf="artificial"),
size_in_gb=0.1,
)

model = TextCrossEncoder("org/model", lazy_load=True, specific_model_path="./")

assert isinstance(model.model, CustomTextCrossEncoder)
assert model.model.model_description.sources.hf == "artificial"


def test_do_not_add_existing_model():
existing_base_model = "sentence-transformers/all-MiniLM-L6-v2"
custom_model_name = "intfloat/multilingual-e5-small"
Expand Down Expand Up @@ -322,22 +337,26 @@ def test_do_not_add_existing_cross_encoder():
existing_base_model = "Xenova/ms-marco-MiniLM-L-6-v2"
custom_model_name = "Xenova/ms-marco-MiniLM-L-4-v2"

with pytest.raises(ValueError, match=f"Model {existing_base_model} is already registered"):
TextCrossEncoder.add_custom_model(
existing_base_model,
sources=ModelSource(hf=existing_base_model),
size_in_gb=0.08,
)
for name in (existing_base_model, existing_base_model.lower()):
with pytest.raises(ValueError, match=f"Model {name} is already registered"):
TextCrossEncoder.add_custom_model(
name,
sources=ModelSource(hf=existing_base_model),
size_in_gb=0.08,
)

TextCrossEncoder.add_custom_model(
custom_model_name,
sources=ModelSource(hf=existing_base_model),
size_in_gb=0.08,
)

with pytest.raises(ValueError, match=f"Model {custom_model_name} is already registered"):
TextCrossEncoder.add_custom_model(
custom_model_name,
sources=ModelSource(hf=custom_model_name),
size_in_gb=0.08,
)
for name in (custom_model_name, custom_model_name.upper()):
with pytest.raises(ValueError, match=f"Model {name} is already registered"):
TextCrossEncoder.add_custom_model(
name,
sources=ModelSource(hf=custom_model_name),
size_in_gb=0.08,
)

assert len(CustomTextCrossEncoder.SUPPORTED_MODELS) == 1
Loading