Transformers from Scratch

AssemblyAI’s free, interactive course Transformers from Scratch shows you how Transformer models operate by developing them from the ground up in PyTorch. Token embeddings, positional encodings, self-attention, multi-head attention, feed-forward networks, encoder/decoder layers, and training loops are among the fundamental building blocks that you will construct yourself rather than depending on high-level libraries. The focus of the course is on comprehending the mathematics and intuition underlying the architecture that drives contemporary LLMs like GPT, BERT, and T5.

Provider/Creator: AssemblyAI
Platform: AssemblyAI Academy
Category: Transformers
Level: Intermediate
Duration: Self-paced
Certificate: No
Rating: ★4.8/5
Direct Course Link: Transformers from Scratch
Recommended For: ML Engineers