Home > AI Solutions > Gen AI > White Papers > Generative AI in the Enterprise with NVIDIA Spectrum-X Networking Platform > Foundation model
Pretraining, fine-tuning, and inferencing large models such as Llama 3 require high performing interconnect to ensure that GPUs in different nodes can communicate efficiently. Based on open standards, the Spectrum-X networking platform provides a scalable networking infrastructure for pretraining and fine-tuning such large models.