LFM2.5-Embedding-350M

Use LFM2.5-Embedding-350M when you need a small, fast vector index or compatibility with standard dense-vector search. Use LFM2.5-ColBERT-350M when retrieval quality matters more than index size.

Specifications

Property	Value
Parameters	~354M
Type	Dense bi-encoder
Document Length	512 tokens
Output	1024-dimensional CLS vector
Similarity	Cosine
Supported Languages	English, Spanish, German, French, Italian, Portuguese, Arabic, Swedish, Norwegian, Japanese, Korean

Semantic Search

Fast dense retrieval for documents and products.

Vector Databases

One vector per item for compact indexing.

Cross-Lingual RAG

Retrieve across 11 supported languages.

Quick Start

sentence-transformers
GGUF

Install:

pip install -U sentence-transformers

Encode queries and documents:

from sentence_transformers import SentenceTransformer

model = SentenceTransformer(
    "LiquidAI/LFM2.5-Embedding-350M",
    trust_remote_code=True,
)

queries = [
    "What is the capital of France?",
    "Which city is Japan's capital?",
]
documents = [
    "Paris is the capital and largest city of France.",
    "Tokyo is the capital of Japan.",
    "Berlin is the capital and largest city of Germany.",
]

query_embeddings = model.encode(
    queries,
    prompt_name="query",
    normalize_embeddings=True,
)
document_embeddings = model.encode(
    documents,
    prompt_name="document",
    normalize_embeddings=True,
)

scores = query_embeddings @ document_embeddings.T
print(scores)

Download GGUF:

hf download LiquidAI/LFM2.5-Embedding-350M-GGUF \
  --local-dir ./LFM2.5-Embedding-350M-GGUF

Use the GGUF files with a llama.cpp build that supports LFM2.5 embedding models.

Models

Fine-tuning

Edge Inference

GPU Inference

Cloud Inference

Guides

Help

LFM2.5-Embedding-350M

Specifications

Semantic Search

Vector Databases

Cross-Lingual RAG

Quick Start

​Specifications

Semantic Search

Vector Databases

Cross-Lingual RAG

​Quick Start

Specifications

Quick Start