NOTE: LoRA models may behave differently to the full model weights.
Dedicated embedding models for semantic search, clustering, classification, and RAG. Send up to 20 inputs per request.