Expand description
Local embedding model backend for Blazen using fastembed.
This crate wraps the fastembed Rust crate (ONNX Runtime) to provide
fully local, offline vector embeddings with no API keys required.
When used through blazen-llm with the fastembed feature flag, this
crate’s FastEmbedModel automatically implements
blazen_llm::EmbeddingModel.
§Quick start
use blazen_embed_fastembed::{FastEmbedModel, FastEmbedOptions};
let model = FastEmbedModel::from_options(FastEmbedOptions::default())?;
let response = model.embed(&["hello".into(), "world".into()]).await?;
assert_eq!(response.embeddings.len(), 2);Structs§
- Fast
Embed Model - A local embedding model backed by
fastembed(ONNX Runtime). - Fast
Embed Options - Options for constructing a
FastEmbedModel. - Fast
Embed Response - Response from a fastembed embedding operation.
Enums§
- Fast
Embed Error - Error type for fastembed operations.