Expand description
Pure Rust embedding via fastembed (ONNX Runtime, no Python).
Downloads model on first use (~25MB for AllMiniLML6V2). Produces 384-dim embeddings in-process with zero external dependencies.
Functions§
- embed
- Embed a single text. Returns 384-dim f32 vector.
- embed_
batch - Embed multiple texts in batch. More efficient than individual calls.
- embedding_
dim - Embedding dimension (AllMiniLML6V2 = 384).