Cerebras WSE (Wafer-Scale Engine)
Realizes: Large-scale deep learning training and inference
The Cerebras Wafer-Scale Engine is a wafer-scale AI accelerator with millions of cores interconnected through a dense on-chip fabric, delivering massive compute for large-scale model training on systems like the CS-2.
Examples
🔗
Cerebras CS-2 system
CS-2 system built around Cerebras WSE-2 to train massive AI models with an entire wafer dedicated to the workload.