Deep dive - Better Attention layers for Transformer models

The self-attention mechanism is at the core of transformer models. As amazing as it is, it requires a significant amount of ...