D2L Julia

A hands-on, interactive deep learning book implemented entirely in Julia, based on the acclaimed Dive into Deep Learning (D2L.ai) textbook. Learn modern machine learning techniques with clean, idiomatic Julia code and a strong emphasis on mathematical intuition.

---

🚀 Features

• 📚 Interactive Notebooks – Every chapter is a live Julia notebook runnable via Pluto.jl or IJulia.jl.
• 🔢 Mathematical Intuition – Clear derivations and explanations alongside the code.
• 🔍 Modern Deep Learning – Covers fundamentals to advanced topics like CNNs, RNNs, Transformers, and beyond.
• 🧪 Built from Scratch – Learn by implementing models and algorithms manually before using frameworks.
• 📦 Flux.jl Powered – Uses the elegant Flux.jl library for building and training models.
• 🧰 Modular Utilities – Reusable utilities for training, data loading, and visualization.

✅ Implemented Chapters

• CH3. Linear Regression
• CH4. Linear Classification
• CH5. MLP
• CH6. Convolutional Neural Networks
• CH7. Modern Convolutional Neural Networks
• CH8. Recurrent Neural Networks
• CH9. Modern Recurrent Neural Networks
• CH10. Attention Mechanisms and Transformers
• CH11. Computer Vision

Getting Started

• Clone the repository:

git clone https://github.com/ashutosh-b-b/d2l-julia.git

• Download Julia via juliaup:

curl -fsSL https://install.julialang.org | sh

• Install Julia 1.11.4

juliaup add 1.11.4

• Instantiate and Precompile the d2lai package:

julia +1.11.4 

julia> using Pkg
julia> Pkg.activate("d2lai/")
julia> Pkg.instantiate()
julia> Pkg.precompile()

🗺️ Roadmap

• Add full coverage for all D2L chapters
• Add automated testing for notebooks
• Add tests for d2lai package.

✍️ Contributing

We welcome contributions!

• Fork this repo
• Open a feature branch
• Submit a pull request (PR)

🙌 Acknowledgments

Original book: Dive into Deep Learning by Aston Zhang, Zachary C. Lipton, Mu Li, and Alex J. Smola

Julia libraries: Flux.jl, Zygote.jl, MLDatasets.jl