Back to Study Plans

35 Problems

Cracking Deep Learning

From perceptrons to transformers. Implement the architectures, loss functions, and training techniques that define modern deep learning.

4 Easy22 Medium9 Hard

Unlock the full study plan

Upgrade to Pro to access the full curriculum.

6 sections

Perceptron Activation Functions Dropout Loss Functions Multi-Layer Perceptron (Forward Pass)Weight Initialization Mini-Batch Training Loop Batch Normalization Backpropagation (Single Hidden Layer)Computational Graph & Autograd Neural Network Foundations Quiz10 questions

Pooling Layers 2D Convolution from Scratch Transposed Convolution VGG Convolution Block Depthwise Separable Convolution LeNet Forward Pass ResNet Residual Block Inception Module Convolutional Networks Quiz10 questions

Vanilla RNN Cell GRU Cell LSTM Cell Sequence-to-Sequence Forward Pass Recurrent Networks Quiz10 questions

Layer Normalization Rotary Position Embedding (RoPE)Scaled Dot-Product Attention Multi-Head Attention Transformer Encoder Block Transformer Decoder Block Vision Transformer Patch Embedding Transformers Quiz10 questions

Autoencoder Forward Pass Variational Autoencoder Forward Pass GAN Training Step Generative Models Quiz10 questions

Non-Maximum Suppression Squeeze-and-Excitation Block U-Net Skip Connection Advanced Architectures & Techniques Quiz10 questions

Summary

Total35

Easy4

Medium22

Hard9

Progress0%

Quizzes0/6