Attention

Deep Learning · advanced · 95 min

A differentiable mechanism for routing information between positions according to learned relevance.

Prerequisites

• Embeddings
• Backpropagation

Unlocks

• Transformers