#1717th of 18 in Learning and cookbooks
annotated_deep_learning_paper_implementations
Annotated PyTorch implementations of deep learning papers with side-by-side notes
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 67.6k
- License
- MIT
- Last commit
- Jan 2026
- Language
- Python
65,442 on 25 Mar 202667,555 on 14 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
A collection of short PyTorch implementations of neural network architectures and algorithms, each rendered on nn.labml.ai as code beside explanatory notes. It covers transformers (including Flash Attention in Triton, RoPE, ALiBi), LoRA, diffusion models, GANs, reinforcement learning, optimizers, normalization layers and more. Installable as the labml-nn pip package.
Who it is for: Engineers and students learning how deep learning papers are implemented
Strengths
- Broad coverage: transformers, diffusion, GANs, RL, optimizers, normalization, LoRA, GNNs, CFR
- Each implementation has side-by-side code and explanation on a hosted website
- Compact PyTorch code that is readable for studying the original papers
- MIT license; install with pip install labml-nn
Weaknesses
- Written for study, not as an optimized production library
- No release tags; versioning is only via the pip package
- Many entries cover older papers; recent architectures are sparse
- README gives no hardware requirements beyond GPU notes for the GPT-NeoX examples
What it needs
- GPU optional
- Needs PyTorch
- Models: PyTorch, JAX (transformer example), Triton (Flash Attention)
Also in Learning and cookbooks
See all 18| Rank | Project | Score |
|---|---|---|
| 1 | ai-engineering-from-scratchHands-on curriculum for building AI systems from math to agents | 83 out of 100 |
| 2 | generative-ai-for-beginners21-lesson course on building generative AI apps in Python and TypeScript | 81 out of 100 |
| 3 | llms-from-scratchCode for the book on building a GPT-style LLM in PyTorch | 76 out of 100 |
| #4 | hello-agentsChinese-language course on building LLM agents from scratch | 76 out of 100 |
| #5 | ml-for-beginners12-week, 26-lesson classic machine learning course using Scikit-learn | 75 out of 100 |