[Learning notes] reading "Attention is all you need" paper

This post contains personal learning notes from reading the "Attention Is All You Need" paper, explaining core concepts like encoder-decoder architecture, attention mechanisms, and the Transformer model's advantages including parallelization and reduced training time. It highlights the paper's breakthrough results on WMT 2014 benchmarks, where a single Transformer model surpassed previous ensemble-based state-of-the-art scores on English-German and English-French translation tasks by significant BLEU margins.