The architectures behind modern AI — CNNs for images, RNNs/LSTM/GRU for sequences, attention and Transformers, NLP tokenization and embeddings, and transfer learning from pretrained models. Every lesson is runnable PyTorch, focused on the architecture itself.
Before you start
This course builds on the Deep Learning with PyTorch course — the tensors and training loop are assumed. You need torch torchvision transformers. A GPU helps but is not required to follow the code.
Convolutional Neural Networks
CNNs are how machines see. A convolution slides a small filter over an image to detect patterns — learn convolution, padding, strides, and pooling from the ground up.
Building a CNN Image Classifier
Assemble convolution, pooling, and dense layers into a working image classifier. See the standard pattern — conv blocks that extract features, then a head that decides.
RNNs, LSTM & GRU
Sequences — text, time series, audio — have order that matters. Recurrent networks process them step by step, carrying a memory; LSTM and GRU fix their short memory.
Attention Mechanisms
RNNs process sequences step by step and forget. Attention lets a model look at every position at once and weigh what matters — the idea that made Transformers possible.
Transformers
The architecture behind modern AI. A Transformer stacks multi-head self-attention and feed-forward layers, processing a whole sequence in parallel — no recurrence.
NLP: Tokenization & Embeddings
Text isn’t numbers. Turn it into tokens, map tokens to learnable embedding vectors, and understand the preprocessing pipeline behind every language model.
Transfer Learning with Pretrained Models
You rarely train from scratch. Start from a model pretrained on millions of examples, adapt it to your task with little data, and get strong results fast.