huggingface/sentence-transformers

Health v2 · maintenance only

Adoption not part of the score

19036 stars · 2867 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Sentence Transformers (SBERT) is a Python library for computing, using, and training state-of-the-art embedding, reranker, sparse encoder, and multi-vector (ColBERT-style) models built on PyTorch and Hugging Face Transformers. It provides access to thousands of pre-trained models on the Hugging Face Hub and supports semantic search, similarity, retrieval, and model finetuning workflows.

Use cases

• compute sentence embeddings for semantic search
• rerank search results with a cross-encoder
• find semantically similar sentences or paraphrases
• build a RAG retrieval pipeline with dense embeddings
• finetune a custom embedding model on my own data
• do ColBERT-style late-interaction retrieval
• cluster or deduplicate documents by meaning
• search images with text using CLIP models

When to choose

• you need text/image embeddings or reranking in Python with minimal code
• you want to train or finetune embedding or reranker models
• you want access to thousands of pre-trained models from the Hugging Face Hub
• you need dense, sparse, or multi-vector retrieval in one library

When to avoid

• you need a production vector database or full search engine rather than an embedding library
• you work outside Python or cannot use PyTorch
• you need lightweight non-neural similarity like TF-IDF or BM25 only

Facets

library · maturity stable

machine-learning nlp rag search-engine image-processing audio-processing video-processing machine-learning deep-learning python cross-platform embeddings sentence-embeddings reranking cross-encoder colbert semantic-search sparse-embeddings transformers pytorch natural-language-processing retrieval-augmented-generation search gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="huggingface/sentence-transformers")