Two 350M retrieval models: one dense, one late interaction
- Source
- x.com
- Date
Introducing LFM2.5-Embedding-350M and LFM2.5-ColBERT-350M: two multilingual retrieval models built for ultra-fast and accurate search across 11 languages. > End-to-end retrieval latency as low as 1.5ms with our enterprise stack! 🚀 > Consistently best-in-class multilingual and cross-lingual performance across Arabic, German, English, Spanish, French, Italian, Japanese, Korean, Norwegian, Portuguese, and Swedish. 🧵
Retrieval models force a tradeoff between speed and accuracy. These two sit at different points on that curve. > LFM2.5-Embedding-350M: one vector per document. Fastest search, smallest index. > LFM2.5-ColBERT-350M: one vector per token. Word-level query matching, higher accuracy, larger index. (2/n)
Both are built on LFM2.5-350M-Base with bidirectional patches, the first full-context encoders in the LFM family. > Both run via llama.cpp GGUFs (CPUs, Laptops, edge devices) <10ms end-to-end query embedding latency. > Both fly on GPUs with our custom runtime <2ms end-to-end query embedding latency. (3/n)
Especially well-suited for short-context collections: product catalogs, FAQ, knowledge bases, support docs, and multilingual search use cases where speed and reliability matter. LFM2.5-Embedding-350M and LFM2.5-ColBERT-350M are available now. > Blog post: https://t.co/OG5MXl6cEW > LFM2.5-Embedding-350M: https://t.co/ZGchsL7Vc7 > LFM2.5-ColBERT-350M: https://t.co/cAZ0cOO99k > Docs: https://t.co/SyQJEpJd6h
Context
Liquid AI's two 350M retrieval models store different representations of searchable text. LFM2.5--350M represents each document with one numerical vector. LFM2.5-ColBERT-350M keeps a vector for each , allowing a query to match individual parts of the document. Release explanation.
That choice changes the search system's storage and matching behavior. The single-vector version targets a smaller index and faster search; the per-token version keeps more detail for matching at the cost of a larger index. Both target multilingual collections such as product catalogs and support documents. The useful comparison is how much matching detail a collection needs relative to the storage and search budget available.
- embedding — A list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.
- token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
postLiquid AI ships LFM2.5-24B-A2B open weights for local agents
postToday we release Open d1: two open-weight multimodal models in our d1 decision…
postLiquid AI's d1 decision model adds vision at a fraction of frontier cost
articled1 is now available on @vercel AI gateway!We worked with the amazing vercel…
Checking sign-in…
Loading comments…