Skip to main content
A quick reference for the built-in models available in Clinia and the situations where each one fits best.
ModelWhen to use
mte-base-knowledge-v1Embed passages with a VECTORIZER processor (see Vectorizer) when you need high-recall semantic search on medical knowledge, guidelines, or FAQs.
mte-base-v1Use in VECTORIZER pipelines for balanced semantic embeddings across general clinical content and structured metadata.
mte-base-knowledge-rank-v1Supply as modelId in search ranking parameters for semantic reranking (see Ranking) to re-order your top-k results without reindexing.
mte-base-sparse-v1Pair with a VECTORIZER step to generate sparse embeddings that preserve keyword and code fidelity for hybrid retrieval workflows.
clinia-segment-v1Run as a SEGMENTER processor (see Segmenter) to split long documents into coherent passages before embedding or retrieval.