Overview of NVIDIA NIM for Table Extraction#
NVIDIA NeMoTM Retriever NIM APIs provide easy access to state-of-the-art models that are foundational building blocks for enterprise semantic search applications, delivering accurate answers quickly at scale. Developers can use these APIs to create robust copilots, chatbots, and AI assistants from start to finish. NeMo Retriever NIM models are built on the NVIDIA software platform, incorporating NVIDIAⓇ CUDAⓇ, NVIDIA TensorRTTM, and NVIDIA TritonTM Inference Server to offer out-of-the-box GPU acceleration. NeMo Retriever includes NIM microservices for creating advanced ingestion and retrieval pipelines, useful for natural language processing (NLP) tasks and generative AI applications like retrieval-augmented generation (RAG).
Ingestion pipelines retrieve documents from external sources beyond the foundational model’s scope. The following NeMo Retriever microservices handle large-scale multimodal document ingestion, detecting, contextualizing, and extracting data from text, tables, charts, and images:
Object Detection NIM - Identifies charts and tables within a PDF.
Table Extraction NIM - Extracts text information from tables, maintaining the reading order of the table.
Chart Extraction NIM - Generates descriptions of charts.
The retrieval pipeline fetches relevant document data and generates responses during inference. The following NeMo Retriever microservices provide superior natural language processing and understanding, boosting retrieval performance:
Text Embedding NIM - Boosts text question-answering retrieval performance, providing high-quality embeddings for many downstream NLP tasks.
Text Reranking NIM - Enhances the retrieval performance further with a fine-tuned reranker, finding the most relevant passages to provide as context when querying an LLM.
The following diagram shows how NeMo Retriever NIM microservices create advanced ingestion and retrieval pipelines for a question-answering RAG application in enterprise settings.
