use-ml-embedding

Embedding vector metadata primitives for RustUse.

Experimental

use-ml-embedding is experimental while use-ml remains below 0.3.0.

Example

use use_ml_embedding::{EmbeddingDimension, EmbeddingModelName, EmbeddingVectorShape};

let model = EmbeddingModelName::new("text-embedding")?;
let dimension = EmbeddingDimension::new(384)?;
let shape = EmbeddingVectorShape::new(dimension);

assert_eq!(model.as_str(), "text-embedding");
assert_eq!(shape.dimension().get(), 384);
# Ok::<(), use_ml_embedding::EmbeddingError>(())

Scope

• Embedding model names, vector IDs, dimensions, vector shapes, modalities, distance metrics, normalization labels, index labels, search labels, and vector formats.
• Metadata only; no embedding computation or vector search.

Non-goals

• RAG pipelines, retrievers, chunkers, document stores, citations, context assembly, or reranking workflows.
• Vector search, embedding computation, model calls, or database clients.

License

Licensed under either Apache-2.0 or MIT.