Create transformers in steps Simple Bigram model Add positional embedding Add self-attention Single Head Add self-attention Multi Head Add Skip connection Add Layer Normalization Add dropouts..... Note: - All the code above are executable from colab notebook.