NVIDIA Nemotron 3 Nano 30B MoE Model Now Available on Amazon SageMaker JumpStart
According to Amazon AWS AI, NVIDIA’s Nemotron 3 Nano 30B model is now generally available in the Amazon SageMaker JumpStart model catalog. The model features 3 billion active parameters and uses a Mixture of Experts (MoE) architecture.
The announcement states that organizations can now access the Nemotron 3 Nano model on Amazon Web Services (AWS) through SageMaker JumpStart, which is designed to help accelerate innovation and deliver business value.
The availability of this model in SageMaker JumpStart allows AWS customers to deploy and utilize NVIDIA’s Nemotron 3 Nano without requiring extensive infrastructure setup. SageMaker JumpStart provides a catalog of pre-trained models that can be quickly deployed for various AI applications.
The MoE architecture with 30B total parameters but only 3B active parameters suggests an efficiency-focused design, potentially offering a balance between model capability and computational requirements.