Learning Transformer Through Implementation

From Multi Head Attention to the Original Transformer model, BERT, and Encoder-Decoder based MarianMT translation model, you'll learn Transformer inside and out by implementing them directly with code.