NVIDIA GH200 Superchip Boosts Llama Model Inference by 2x

The NVIDIA GH200 Grace Hopper Superchip accelerates inference on Llama models by 2x, enhancing user interactivity without compromising system throughput, according to NVIDIA.