Liquid AILiquid Nanos
LFM2.5-Embedding-350M
LFM2.5-Embedding-350M creates one vector per text for multilingual retrieval.
What the release provides
Its bidirectional backbone produces a 1,024-dimensional CLS vector compared with cosine similarity. The approximately 354M model targets retrieval across 11 languages and uses the LFM Open License.LFM2.5-Embedding-350M official model card
How to use it well
Practical guidance. Rebuild the document index when changing embedding models. Evaluate query-document relevance across the actual language pairs and document lengths.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | LiquidAI/LFM2.5-Embedding-350M |