Infrastructure & Agents

What Is Distributed Training?

Distributed training coordinates many GPUs or servers so they can jointly train a model that would be too large or slow for a single device. It relies on strategies like data, model, and pipeline parallelism to share the work.