Skip to content
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

feat (v2): Update so model_max_length updates max_seq_length for Sentence Transformers #8334

Merged
merged 4 commits into from
Sep 6, 2024
Merged

feat (v2): Update so model_max_length updates max_seq_length for Sentence Transformers #8334

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
3286597
Update so model_max_length does what is expected
sjrl Sep 5, 2024
04f781f
Add release notes
sjrl Sep 5, 2024
e0c4444
Some fixes
sjrl Sep 5, 2024
17a722c
Another test
sjrl Sep 5, 2024
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
2 changes: 2 additions & 0 deletions haystack/components/embedders/sentence_transformers_document_embedder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -187,6 +187,8 @@ def warm_up(self):
model_kwargs=self.model_kwargs,
tokenizer_kwargs=self.tokenizer_kwargs,
)
if self.tokenizer_kwargs and self.tokenizer_kwargs.get("model_max_length"):
self.embedding_backend.model.max_seq_length = self.tokenizer_kwargs["model_max_length"]

@component.output_types(documents=List[Document])
def run(self, documents: List[Document]):
Expand Down
2 changes: 2 additions & 0 deletions haystack/components/embedders/sentence_transformers_text_embedder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,6 +173,8 @@ def warm_up(self):
model_kwargs=self.model_kwargs,
tokenizer_kwargs=self.tokenizer_kwargs,
)
if self.tokenizer_kwargs and self.tokenizer_kwargs.get("model_max_length"):
self.embedding_backend.model.max_seq_length = self.tokenizer_kwargs["model_max_length"]

@component.output_types(embedding=List[float])
def run(self, text: str):
Expand Down
4 changes: 4 additions & 0 deletions releasenotes/notes/update-max-seq-lenght-st-1dc3d7a9c9a3bdcd.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
---
features:
- |
Updates SentenceTransformersDocumentEmbedder and SentenceTransformersTextEmbedder so model_max_length passed through tokenizer_kwargs also updates the max_seq_length of the underly SentenceTransformer model.
8 changes: 6 additions & 2 deletions test/components/embedders/test_sentence_transformers_document_embedder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -226,18 +226,22 @@ def test_from_dict_none_device(self):
)
def test_warmup(self, mocked_factory):
embedder = SentenceTransformersDocumentEmbedder(
model="model", token=None, device=ComponentDevice.from_str("cpu")
model="model",
token=None,
device=ComponentDevice.from_str("cpu"),
tokenizer_kwargs={"model_max_length": 512},
)
mocked_factory.get_embedding_backend.assert_not_called()
embedder.warm_up()
embedder.embedding_backend.model.max_seq_length = 512
mocked_factory.get_embedding_backend.assert_called_once_with(
model="model",
device="cpu",
auth_token=None,
trust_remote_code=False,
truncate_dim=None,
model_kwargs=None,
tokenizer_kwargs=None,
tokenizer_kwargs={"model_max_length": 512},
)

@patch(
Expand Down
10 changes: 8 additions & 2 deletions test/components/embedders/test_sentence_transformers_text_embedder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -201,17 +201,23 @@ def test_from_dict_none_device(self):
"haystack.components.embedders.sentence_transformers_text_embedder._SentenceTransformersEmbeddingBackendFactory"
)
def test_warmup(self, mocked_factory):
embedder = SentenceTransformersTextEmbedder(model="model", token=None, device=ComponentDevice.from_str("cpu"))
embedder = SentenceTransformersTextEmbedder(
model="model",
token=None,
device=ComponentDevice.from_str("cpu"),
tokenizer_kwargs={"model_max_length": 512},
)
mocked_factory.get_embedding_backend.assert_not_called()
embedder.warm_up()
embedder.embedding_backend.model.max_seq_length = 512
mocked_factory.get_embedding_backend.assert_called_once_with(
model="model",
device="cpu",
auth_token=None,
trust_remote_code=False,
truncate_dim=None,
model_kwargs=None,
tokenizer_kwargs=None,
tokenizer_kwargs={"model_max_length": 512},
)

@patch(
Expand Down