Sentence Transformers

Sentence Transformers

Sentence Transformers is the go-to Python library for computing dense sentence and text embeddings. It provides over 5,000 pretrained models on the Hugging Face Hub and a simple API for computing embeddings used in semantic search, sentence similarity, clustering, information retrieval, and RAG pipelines.

Key Features

• 5,000+ pretrained models: all-MiniLM, BAAI/bge, E5, GTE, multilingual models, and more
• Semantic textual similarity: cosine similarity for sentence pairs
• Semantic search: bi-encoder retrieval with FAISS, Annoy, or built-in search
• Clustering: K-Means, Agglomerative clustering on embeddings
• Training: contrastive learning (CoSENT, MultipleNegativeRankingLoss) to fine-tune on custom data
• Multi-lingual: 50+ language support

Quick Start

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("all-mpnet-base-v2")

sentences = ["The cat sat on the mat", "A feline rested on a rug"]
embeddings = model.encode(sentences)

# Semantic similarity
from sentence_transformers import util
similarity = util.cos_sim(embeddings[0], embeddings[1])
print(f"Similarity: {similarity.item():.3f}")  # 0.878

Install via ai-supply

npx ai-supply add sentence-transformers-embeddings

Curated mirror of the open-source Sentence Transformers (Apache-2.0). Get it from the source.