Custom Algorithm-based Fault Tolerance for Attention Layers in Transformers

Open in new window