A 3-hour hands-on workshop. We'll build a Transformer architecture from scratch in PyTorch: attention, masking, positional encoding and the training loop.
Anyone familiar with PyTorch who wants to get into NLP. Linear algebra and basic ML knowledge is enough.