NVIDIA has made it easier to train AI models using its H100 GPUs in the cloud. The company's DGX Cloud platform now supports training with these powerful chips, allowing users to access high-performance computing resources without the need for on-premises hardware. This move is aimed at accelerating AI research and development by providing a scalable and cost-effective solution for model training. --- Why it matters: This matters because it gives researchers and engineers more flexibility in their work, enabling them to focus on developing and testing models rather than managing infrastructure. By accessing powerful GPUs in the cloud, they can train larger models faster and with less hassle. Source: https://huggingface.co/blog/train-dgx-cloud