v1.0 — Now covering modern LLM architectures
Deep Learning,documented properly.
A concise, opinionated reference for the concepts that matter — from the perceptron to the transformer. Built for engineers who want depth without the dissertation.
Browse topics
6 conceptsNeural Networks
Perceptrons, MLPs, and the universal approximation theorem.
Read
Backpropagation
The chain rule, gradients, and how networks actually learn.
Read
Optimization
SGD, momentum, Adam, and learning rate scheduling.
Read
CNNs
Convolutions, pooling, and translation invariance.
Read
RNNs & LSTMs
Sequence modeling, gates, and vanishing gradients.
Read
Transformers
Self-attention, positional encodings, and scaling laws.
Read
Why another deep learning reference?
Most resources are either textbook-dense or surface-level blog posts. DeepDocs sits in the middle — explaining each concept with the right level of math, intuition, and code, all in one place.