109-language shared space
Hub data records sentence-similarity with multilingual tags, and the model card states the port maps 109 languages to a shared vector space.
Open Source Model Profile · sentence-transformers
LaBSE
LaBSE is a 471M-parameter BERT embedding model from sentence-transformers. According to the model card, it maps 109 languages to a shared vector space.
- Publisher
- sentence-transformers
- Task
- sentence-similarity
- Model type
- bert
- License
- apache-2.0
- Library
- sentence-transformers
- Publication status
- Accepted · not indexed
Model overview
LaBSE is published by sentence-transformers as a BERT sentence-similarity model. The captured configuration identifies BertModel with model type bert, and Safetensors metadata reports 470,927,360 parameters. According to the model card, it is a PyTorch port mapping 109 languages to a shared vector space.
Recorded capabilities
BERT 768-dimension embeddings
Captured configuration records BertModel with about 471M parameters, and the card documents 768-dimension CLS pooling with Dense Tanh and normalization.
Documented SentenceTransformer use
According to the model card, use is via SentenceTransformer with model.encode, max sequence length 256, and lower-casing disabled.
Apache-2.0 licensing record
Card data records apache-2.0 for this model.
Use cases in the source record
- Multilingual sentence similarity and retrieval workflows that encode sentences to shared-space vectors with model.encode.
- Embedding-pipeline integration using the documented 768-dimension normalized output and 256-token maximum sequence.
Limitations and unknowns
- No context-window value was extracted from this record.
- Language coverage and publication claims come from the publisher model card and cited LaBSE publication without independent verification here.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
Source and provenance
Source: sentence-transformers/LaBSE
Captured: Unknown. Processed: 2026-09-07T19:34:57.668635+00:00.
LaBSE This is a port of the LaBSE model to PyTorch. It can be used to map 109 languages to a shared vector space. Usage (Sentence-Transformers) Using this model becomes easy when you have sentence-transformers installed: pip install -U sentence-transformers Then you can use the model like this: from sentence_transformers import SentenceTransformer sentences = [ "This is an example sentence" , "Each sentence is converted" ] model = SentenceTransformer( 'sentence-transformers/LaBSE' ) embeddings = model.encode(sentences) print (embeddings) Full Model Architecture SentenceTransformer( (0): Transformer({'max_seq_length': 256, 'do_lower_…
F001F002F003F004F005F006F007F008F009F010F011F012F013