frankensearch-embed-0.1.0 has been yanked.
Embedder implementations for the frankensearch hybrid search library.
Provides three tiers of text embedding:
- Hash (
hashfeature, default): FNV-1a hash embedder, zero dependencies, always available. Model2Vec(model2vecfeature): potion-128M static embedder, fast tier (~0.57ms).FastEmbed(fastembedfeature): MiniLM-L6-v2 ONNX embedder, quality tier (~128ms).
The EmbedderStack auto-detection probes for available models and configures
the best fast+quality pair automatically.