Andrej Karpathy @karpathy ↗

Run the full 2.78-trillion-parameter Kimi K3 model, DeepSeek V4.1 Flash or GLM-5.3-Flash beyond available RAM by streaming activated weights directly from NVMe. A dependency-free, embeddable C inference engine.

Multi-subscription extension for pi -- use multiple OAuth accounts per provider (Anthropic, Codex, Copilot, Gemini, Antigravity)

Very low latency speech to text, intent recognition, and text to speech, for building voice agents and interfaces

Beautiful, Modern & Opinionated Linux

Minimal Claude Code alternative. Single Python file, zero dependencies, ~250 lines.

The open source coding agent.

A tiny CPU simulator written in Python

Our library for RL environments + evals

FlexAttention based, minimal vllm-style inference engine for fast Gemma 2 inference.

gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI

Wan: Open and Advanced Large-Scale Video Generative Models

Simple MPI implementation for prototyping or learning

The simplest, fastest repository for training/finetuning small-sized VLMs.

Minimalistic 4D-parallelism distributed training framework for education purpose

Implementing DeepSeek R1's GRPO algorithm from scratch

Frontier Models playing the board game Diplomacy.

Minimal reproduction of DeepSeek R1-Zero

Official repository for our work on micro-budget training of large-scale diffusion models.

NanoGPT (124M) in 90 seconds

A PyTorch native platform for training generative AI models

Efficient Triton Kernels for LLM Training

Apple MLX native implementations of state-of-the-art generative image & video models

Official inference repo for FLUX.1 models

the scott CPU from "But How Do It Know?" by J. Clark Scott

Run PyTorch LLMs locally on servers, desktop and mobile

Animation engine for explanatory math videos

A lightweight library for portable low-level GPU computation using WebGPU.

Simple Byte pair Encoding mechanism used for tokenization process . written purely in C

UNet diffusion model in pure CUDA

[ICLR 2025] Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling

gpt-2 from scratch in mlx

User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

Implementation of Diffusion Transformer (DiT) in JAX

Implementation for MatMul-free LM.

Schedule-Free Optimization in PyTorch

The AI Compute Platform for frontier teams. SkyPilot turns fragmented AI compute into one AI supercomputer, so frontier AI teams build custom intelligence faster.

Vane is an AI-powered answering engine.

My favorite C programming practices.

Tile primitives for speedy kernels

A minimal GPU design in Verilog to learn how GPUs work from the ground up

lightweight, standalone C++ inference engine for Google's Gemma models.

LLM inference in C/C++

Distribute and run LLMs with a single file.

A subset of PyTorch's neural network modules, written in Python using OpenAI's Triton.

Fast bare-bones BPE for modern tokenizer training

The official PyTorch implementation of Google's Gemma models

A benchmark to evaluate language models on questions I've previously asked them to solve.

Official implementation for the paper: "Code Generation with AlphaCodium: From Prompt Engineering to Flow Engineering""

A high-throughput and memory-efficient inference and serving engine for LLMs

MLX: An array framework for Apple silicon

RuLES: a benchmark for evaluating rule-following in language models

Code and documentation to train Stanford's Alpaca models, and generate the data.

Fine-tune mistral-7B on 3090s, a100s, h100s

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4, MiniMax-H3, Gemma 4, FLUX and more.

A python script to help manage a Gmail inbox by filtering out promotional emails using GPT-3 or GPT-4.

Master programming by recreating your favorite technologies from scratch.

Xv6 for RISC-V

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

The TinyLlama project is an open endeavor to pretrain a 1.1B Llama model on 3 trillion tokens.

Implementation of the LLaMA language model based on nanoGPT. Supports flash attention, Int8 and GPTQ 4bit quantization, LoRA and LLaMA-Adapter fine-tuning, pre-training. Apache 2.0-licensed.

OpenAssistant is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so.

Train to 94% on CIFAR-10 in <6.3 seconds on a single A100. Or ~95.79% in ~110 seconds (or less!)

Tesseract Open Source OCR Engine (main repository)

The agent engineering platform.

LlamaIndex is the document processing platform for AI

Cramming the training of a (BERT-type) language model into limited compute.

Hackable and optimized Transformers building blocks, supporting a composable construction.

Robust Speech Recognition via Large-Scale Weak Supervision

🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.

Official implementation of VQ-Diffusion

Development repository for the Triton language and compiler

JAX implementation ViT-VQGAN

DALL·E Mini - Generate images from a text prompt

min(DALL·E) is a fast, minimal port of DALL·E Mini to PyTorch

Fast and memory-efficient exact attention

Repository for the book

Version control for machine learning

[ICCV 2019] Monocular depth estimation from a single image

StudioGAN is a Pytorch library providing implementations of representative Generative Adversarial Networks (GANs) for conditional/unconditional image generation.

Taming Transformers for High-Resolution Image Synthesis

[NeurIPS‘2021] "TransGAN: Two Pure Transformers Can Make One Strong GAN, and That Can Scale Up", Yifan Jiang, Shiyu Chang, Zhangyang Wang

Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch

Flexible and powerful tensor operations for readable and reliable code (for pytorch, jax, TF and others)

The PyTorch Implementation based on YOLOv4 of the paper: "Complex-YOLO: Real-time 3D Object Detection on Point Clouds"

End-to-End Object Detection with Transformers

Image augmentation for machine learning experiments.

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

The largest collection of PyTorch image encoders / backbones. Including train, eval, inference, export scripts, and pretrained weights -- ResNet, ResNeXT, EfficientNet, NFNet, Vision Transformer (ViT), MobileNetV4, MobileNet-V3 & V2, RegNet, DPN, CSPNet, Swin Transformer, MaxViT, CoAtNet, ConvNeXt, and more

Latex code for making neural networks diagrams

A pytorch implementation of Detectron. Both training from scratch and inferring directly from pretrained Detectron weights are available.

Fast math typesetting for the web.

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Supporting public code for SIGBOVIK17 submission

Automatically track which applications you use and for how long.

A beginner's guide to setting up a development environment on macOS

Dense image captioning in Torch

Javascript neural networks on top of general scalar/tensor reverse-mode automatic differentiation.

Deep Networks with Stochastic Depth

Efficient, reusable RNNs and LSTMs for torch

Torch implementation of ResNet from http://arxiv.org/abs/1512.03385 and training scripts

From the basics to slightly more interesting applications of Tensorflow

Deep Residual Learning for Image Recognition

GPU Powered BLAS for Browsers :gem:

Adversarial networks in TensorFlow

A parser for Google Scholar, written in Python

Multi-layer Recurrent Neural Networks (LSTM, RNN) for character-level language models in Python using Tensorflow

Simple tutorials using Google's TensorFlow Framework

Marvin: A Minimalist GPU-only N-Dimensional ConvNets Framework

An Open Source Machine Learning Framework for Everyone

Faster R-CNN

The open source frontend for GitBook doc sites

Torch implementation of neural style algorithm

Neural Slime Volleyball

Bare bones introduction to machine learning from linear regression to convolutional neural networks using Theano.

DeepWalk - Deep Learning for Graphs

Implements SFO minibatch optimizer in Python and MATLAB, and reproduces figures from paper.

A microframework for working with Amazon's Mechanical Turk

Deep Pink is a chess AI that learns to play chess using deep learning.

Benchmark testbed for assessing the performance of optimisation algorithms

The SDK for Jetpac's iOS Deep Belief image recognition framework

JavaScript GIF encoding library

:globe_with_meridians: Jekyll is a blog-aware static site generator in Ruby

All stories and comments posted on Hacker News upto May 29, 2014

A basic implementation of convolutional neural nets

A Python-embedded modeling language for convex optimization problems.

A JavaScript library aimed at visualizing graphs of thousands of nodes and edges

BING Objectness proposal estimator linux/mac/windows version implementation, runs at 1000 FPS. More in http://mmcheng.net/bing/ and also http://www.robots.ox.ac.uk/~szheng/DepthProposals.html

C-based/Cached/Core Computer Vision Library, A Modern Computer Vision Library

R-CNN: Regions with Convolutional Neural Network Features

Pure Python version of the mlabwrap Python to Matlab bridge

A Pythonic wrapper for the Wikipedia API

Google Mirror API's Quickstart for Python

Teardown of Google Glass

JavaScript with Batteries Included for Google Glass

JavaScript Computer Vision library.

Random Forest implementation for JavaScript. Supports arbitrary weak learners. Includes interactive demo.

Support Vector Machine in Javascript (SMO algorithm, supports arbitrary kernels) + GUI demo

Artificial Life simulator using canvas. Based on https://github.com/karpathy/scriptsbots

Artificial Life simulator using canvas. Based on https://github.com/karpathy/scriptsbots

Face detection for cats in JavaScript - demo for TXJS 2012 talk

[UNMAINTAINED] Histogram of Oriented Gradients (HOG) descriptor extractor

A wrapper for LibSVM that lets you train SVM's directly on Eigen library matrices in C++

A Random Forest implementation for MATLAB. Supports arbitrary weak learners that you can define.

####### ALERT! #########: my fork of the project has moved:

Game making library for using Canvas element

Automating research publications discovery and analysis. For example, ever wish your computer could automatically open papers that are most similar to a paper at an arbitrary url? How about finding all papers that report results on some dataset? Let's re-imagine literature review.

Matlab/Octave toolbox for deep learning. Includes Deep Belief Nets, Stacked Autoencoders, Convolutional Neural Nets, Convolutional Autoencoders and vanilla Neural Nets. Each method has examples to get you started.

Google+ Blog

Numerical analysis in Javascript

ScriptBots is an Open Source Evolutionary Artificial Life Simulation of Predator-Prey dynamics, written by Andrej Karpathy.

Ensemble of Exemplar-SVMs for Object Detection and Beyond

useful scripts to work with Twitter + Python. Requires the tweepy library.

ScriptBots is an Open Source Evolutionary Artificial Life Simulation of Predator-Prey dynamics, written by Andrej Karpathy.

ScriptBots is an Open Source Evolutionary Artificial Life Simulation of Predator-Prey dynamics, written by Andrej Karpathy.

Defunct

scikit-learn: machine learning in Python

Simple feed-forward neural network in JavaScript

JavaScript 3D Library.

Code repository for jQuery Fundamentals training

Twitter for Python!