NVIDIA launches Ising open quantum AI models for calibration and error correction

NVIDIA released Ising, which it calls a family of open quantum AI models. Ising Calibration is a vision-language model that reads quantum processor measurements, and Ising Decoding is a 3D convolutional network for real-time quantum error correction. The company's own figures claim calibration time drops from days to hours and decoding is up to 2.5x faster and 3x more accurate than PyMatching.

• Ising Calibration: AI-driven quantum processor calibration, days to hours (NVIDIA claim)
• Ising Decoding: up to 2.5x faster and 3x more accurate than PyMatching (NVIDIA claim)
• Available on GitHub, Hugging Face and build.nvidia.com; license terms not stated
• Adopters named include IonQ, IQM, Infleqtion, Q-CTRL and Fermi National Accelerator Laboratory