Liquid AI releases two 350M parameter retrieval models

🤖 Liquid AI releases two 350M parameter retrieval models

Liquid AI has introduced multilingual models LFM2.5-Embedding-350M and LFM2.5-ColBERT-350M. These are the first bidirectional models in the LFM family, optimized for fast retrieval with support for 11 languages.

🌍 The emergence of efficient and compact (350M) encoders allows for significantly reducing costs and latency when deploying RAG systems and search engines.

👤 High-quality search and FAQ systems can now be run locally on powerful laptops with very high speeds.

Source 1: https://www.liquid.ai/blog/lfm2-5-retrievers