intfloat/e5-large-v2
🧠 AI Modelintfloat
High-quality sentence embedding model for semantic similarity and retrieval tasks.
e5-large-v2 is a sentence-transformer model fine-tuned from BERT-large with a contrastive learning objective on a diverse corpus of web and Wikipedia text. It outputs 1024-dimensional vectors and supports both symmetric (e.g., semantic similarity) and asymmetric (e.g., query-document) retrieval tasks. The model is optimized for efficiency with ONNX, OpenVINO, and SafeTensors support. It is one of the most downloaded models on HuggingFace for sentence embeddings, with over 2.5 million downloads. Its strong performance on the MTEB leaderboard makes it a go-to choice for production embedding needs.
💡Highlights
- ├─State-of-the-art on MTEB benchmark
- ├─1024-dim embeddings, BERT-large backbone
- └─2.5M+ downloads, open-source
🎯For
- ├─NLP developers
- ├─search engineers
- └─ML researchers