Thermodynamic Hypergraphical Model Library in JAX

A multi-cloud framework for big data analytics and embarrassingly parallel jobs, that provides an universal API for building parallel applications in the cloud ☁️🚀

Unofficial revival of the well known .NET debugger and assembly editor, dnSpy

SoTA Approach for ARC-AGI 2

ACI.dev is the open source tool-calling platform that hooks up 600+ tools into any agentic IDE or custom AI agent through direct function calling or a unified MCP server. The birthplace of VibeOps.

Domain Specific Language for the Abstraction and Reasoning Corpus

Three-stage pipeline for generating MCP servers on demand. Built with Google ADK agents and FastMCP.

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

The Abstraction and Reasoning Corpus

Hierarchical Reasoning Model Official Release

bluetooth mesh chat, IRC vibes

Official PyTorch implementation for "Large Language Diffusion Models"

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

CIFAR-10 speedruns: 94% in 2.6 seconds and 96% in 27 seconds

NanoGPT (124M) in 3 minutes

Fully open reproduction of DeepSeek-R1

Code samples for my book "Neural Networks and Deep Learning"

A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API

Code samples for my book "Neural Networks and Deep Learning"

You like pytorch? You like micrograd? You love tinygrad! ❤️

Aidan Bench attempts to measure <big_model_smell> in LLMs.

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek-R1, Qwen3, Gemma 3, TTS 2x faster with 70% less VRAM.

Data and tools for generating and inspecting OLMo pre-training data.

Modeling, training, eval, and inference code for OLMo

Attempt to make multiple residual streams from Bytedance's Hyper-Connections paper accessible to the public

Testing baseline LLMs performance across various models