Attention 101

From dot products to multi-head transformers. Each step composes onto the next.

1. 1. ○ Implement Scaled Dot-Product Attention
2. 2. ○ Self-Attention Layer
3. 3. ○ Multi-Query Attention
4. 4. ○ Transformer Encoder Block
5. 5. ○ Multi-Head Attention Block