Skip to content

Phase 2: The Architectural Era

Goal: Master spatial and temporal data processing.

Modules

Module Status Description
Convolutions 🔲 Conv2d, im2col, dilated convolutions
Pooling 🔲 MaxPool, AvgPool, GlobalPool
Skip Connections 🔲 Residual blocks
CNN Architecture 🔲 Full ResNet-18 implementation
Vanilla RNN 🔲 RNN cell, BPTT
LSTM & GRU 🔲 Gating mechanisms
Seq2Seq 🔲 Encoder-decoder, teacher forcing

🎯 Capstone

Train ResNet-18 on CIFAR-10 and a character-level LSTM language model.