NVIDIA Releases Nemotron-Labs-Sigma-distill-3B-nfe2 Model

NVIDIA has launched the Nemotron-Labs-Sigma-distill-3B-nfe2, a 3 billion-parameter continuous diffusion language model.

NVIDIA has released the Nemotron-Labs-Sigma-distill-3B-nfe2, a 3 billion-parameter language model designed for fast few-step text generation using Phased Distribution Matching Distillation. This model, part of the Nemotron-Labs-Sigma architecture family, features a maximum context length of 4,096 tokens and is optimized for latency-sensitive generation tasks. It is available for both commercial and non-commercial use under the NVIDIA Nemotron Open Model License. The model runs offline on user hardware, specifically requiring a CUDA-capable NVIDIA GPU, as CPU inference is not supported. According to the model card, NVIDIA claims that the model achieves faster training and inference times by leveraging NVIDIA’s GPU-accelerated systems and software frameworks, such as CUDA libraries.