NeMo Curator

NeMo Curator is a scalable, GPU-accelerated data curation Microservices that simplifies data curation tasks such as text extraction, cleaning, and deduplication.

NeMo Customizer

NeMo Customizer is Microservices that simplifies fine-tuning and alignment of LLM for domain-specific use cases. It currently supports LoRA and p-turning, and other alignment methods such as SFT will be supported in the future. (As of May 1, 2024)

NeMo Evaluator

NeMo Evaluator is Microservices that evaluates custom-generated AI models using various academic and custom benchmarks. By using it in conjunction with NeMo Customizer, you can continuously customize and evaluate your models to improve their performance.

NeMo Retriever

NeMo Retriever is Microservices that performs tasks such as retrieving data such as PDFs, encoding them, and searching for related information when using RAG (Search Augmentation Generation). It is built with NVIDIA​ ​SDKs such as CUDA and NVIDIA Tensor RT, and is easy to use and has excellent performance.

NeMo Guardrails

NeMo Guardrails is a toolkit that helps suppress the creation of answers on undesirable topics and misleading content in generative AI applications, helping to develop secure and reliable LLM systems.

NVIDIA NIM

NVIDIA NIM is part of NVIDIA AI Enterprise, an inference Microservices designed to accelerate the deployment of generative AI in the enterprise.

Supporting open source models, NVIDIA AI Foundation models and more, NIM provides industry-standard APIs to enable seamless and scalable AI inference on-premise or in the cloud.