NVIDIA/NeMo-Retriever
NVIDIA's document extraction and embedding pipeline for retrieval-augmented generation and generative AI applications.
Velocity · 7d
+4.5
★ / day
Trend
→steady
star history
NeMo Retriever Library extracts text, tables, charts, and infographics from documents using OCR and classification, then computes vector embeddings for the extracted content and stores them in LanceDB for downstream generative AI and RAG applications. It leverages NVIDIA NIM microservices for scalable, production-grade document processing and can be deployed on Kubernetes using Helm charts.