Neural Networks: Zero to Hero is Andrej Karpathy’s educational course repository for learning neural networks from the ground up. It accompanies a series of video lectures where learners code and train neural networks step by step. The project starts from basic concepts and gradually builds toward more capable character-level language models. Its notebooks are written to make the mechanics visible, so learners can understand gradients, backpropagation, embeddings, multilayer perceptrons, batch normalization, and Transformer-style ideas through code. The course is especially valuable because it avoids treating neural networks as black boxes. nn-zero-to-hero is best suited for programmers who want a practical, code-first introduction to how modern neural networks work.
Features
- Jupyter notebooks from the lecture series
- Code-first neural network instruction
- Backpropagation and gradient learning examples
- Character-level language modeling projects
- Exercises linked to individual lectures
- Practical foundation for understanding modern LLMs
Project Samples
