NVIDIA

A collection of speculative decoding modules created using TensorRT Model Optimizer.

Framework of PyTorch composable modules for developing physics guided machine learning training pipelines. https://github.com/NVIDIA/physicsnemo

Ultra-efficient reasoning model! SOTA Accuracy / CoT Length trade-offs

Collection of RL verifiable data for NeMo Gym

Reasoning data for supervised finetuning of LLMs to advance data distillation for competitive coding

NV-Embed is a generalist embedding model encompassing retrieval, reranking, classification, clustering, STS tasks.

BigVGAN is a universal neural vocoder that generates audio waveform using mel spectrogram as input.

Enabling 4k resolution for VLMs, CVPR 2025, https://nvlabs.github.io/PS3/

A collection of generative models quantized and optimized for inference with TensorRT Model Optimizer.

NVIDIA Clara Models for Biology

3D-Informed World-Consistent Video Generation with Precise Camera Control

Collection of open, commercial-grade datasets for physical AI developers

Eagle is a family of frontier vision-language models with data-centric strategies. The model supports both HD image and long-context video input.

A family of frontier-class multimodal large language models (LLMs) that achieve state-of-the-art results on vision-language tasks and text-only tasks.

NeMo Parakeet ASR Models attain strong speech recognition accuracy while being efficient for inference. Available in CTC and RNN-Transducer variants.

The Nemotron 3 8B Family of models is optimized for building production-ready generative AI applications for the enterprise.

A family of compressed models obtained via pruning and knowledge distillation