Embedding Models
Embedding models transform text into high-dimensional vectors that represent semantic meaning. The quality of these embeddings is crucial for effective retrieval.
Key Considerations:
-
Must support both the query and the document chunks.
-
Should be trained or fine-tuned on similar types of text.
-
Trade-off: Smaller models are faster; larger models are more accurate.
Common Embedding Models:
| Model Name | Type | Notes |
|---|---|---|
| intfloat/e5-small-v2 | Dense | Lightweight and fast; good baseline |
| intfloat/e5-large-v2 | Dense | More accurate; used for robust retrieval |
| BAAI/bge-small-en | Dense | High-quality general-purpose model |
| BAAI/bge-large-en-v1.5 | Dense | One of the most accurate public models |
| GTE-base / GTE-small | Dense | Lightweight alternatives by Alibaba |
| all-mpnet-base-v2 | Dense | From SentenceTransformers; widely used baseline |
| Jina Embeddings v2 | Dense | Optimized for semantic search |
Choosing a Model:
- for production use we went with
nomic-embed-text.