NVIDIA Launches Open-Source NIXL Library to Speed AI Inference Data Transfers

NVIDIA releases Inference Transfer Library (NIXL), an open-source tool accelerating KV cache transfers for distributed AI inference across major cloud platforms. (Read More)