How huggingface/sentence-transformers works
Sentence Transformers is a Python framework for producing dense, sparse, and token-level embeddings, scoring text pairs with rerankers, and training or evaluating these model families. The principal workflow is user input to a model’s encode or predict operation, then embeddings or scores; similarity and retrieval utilities support downstream search tasks. The graph also preserves the documented training and evaluation branches. Relationships are based on the README and supplied source excerpts/index; unsampled implementation details are not inferred.
Main entry points: sentence_transformer/model.py SentenceTransformer · cross_encoder/model.py CrossEncoder · sparse_encoder/model.py SparseEncoder · multi_vector_encoder/model.py MultiVectorEncoder
Components 21
Dense embeddings
- SentenceTransformer
sentence_transformers/sentence_transformer/model.py - Dense embeddings
- Similarity functions
sentence_transformers/util/similarity.py - Retrieval utilities
sentence_transformers/util/retrieval.py
Cross-encoder reranking
- CrossEncoder
sentence_transformers/cross_encoder/model.py - Pair scores
Sparse embeddings
- SparseEncoder
sentence_transformers/sparse_encoder/model.py - Sparse embeddings
- Sparse search engines
sentence_transformers/sparse_encoder/search_engines.py
Multi-vector retrieval
- MultiVectorEncoder
sentence_transformers/multi_vector_encoder/model.py - Token embeddings
- Late-interaction scoring
sentence_transformers/multi_vector_encoder/scoring/colbert.py
Training and evaluation
- SentenceTransformer trainer
sentence_transformers/sentence_transformer/trainer.py - CrossEncoder trainer
sentence_transformers/cross_encoder/trainer.py - SparseEncoder trainer
sentence_transformers/sparse_encoder/trainer.py - MultiVectorEncoder trainer
sentence_transformers/multi_vector_encoder/trainer.py - Embedding and retrieval evaluators
sentence_transformers/sentence_transformer/evaluation/information_retrieval.py - Reranking evaluators
sentence_transformers/cross_encoder/evaluation/reranking.py - Sparse evaluators
sentence_transformers/sparse_encoder/evaluation/sparse_information_retrieval.py - Multi-vector evaluators
sentence_transformers/multi_vector_encoder/evaluation/information_retrieval.py
Other components
- Application user
Connections 20
- Application user → SentenceTransformer: encodes text
README.md - SentenceTransformer → Dense embeddings: returns embeddings
README.md - Application user → Similarity functions: compares vectors
README.md - Application user → Retrieval utilities: runs search
README.md - Application user → CrossEncoder: scores text pairs
README.md - CrossEncoder → Pair scores: returns scores
README.md - Application user → SparseEncoder: encodes text
README.md - SparseEncoder → Sparse embeddings: returns embeddings
README.md - Application user → Sparse search engines: searches sparse vectors
- Application user → MultiVectorEncoder: encodes query and document
README.md - MultiVectorEncoder → Token embeddings: returns token vectors
README.md - Application user → Late-interaction scoring: scores late interaction
README.md - Application user → SentenceTransformer trainer: trains dense model
README.md - Application user → CrossEncoder trainer: trains reranker
README.md - Application user → SparseEncoder trainer: trains sparse model
README.md - Application user → MultiVectorEncoder trainer: trains multi-vector model
README.md - Application user → Embedding and retrieval evaluators: evaluates retrieval
README.md - Application user → Reranking evaluators: evaluates reranking
README.md - Application user → Sparse evaluators: evaluates sparse retrieval
README.md - Application user → Multi-vector evaluators: evaluates retrieval
README.md