[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

A curated collection of public industrial datasets.

From babyGPT to diffusion GPT: An annotated implementation of a character-level discrete diffusion model (adapted from Karpathy’s baby GPT).

Python GUI builder. GUI builder for Tkinter, CustomTkinter, Kivy and PySide (upcoming)

A modular framework for neural networks with Euclidean symmetry

Optimize prompts, code, and more with AI-powered Reflective Text Evolution

Training-Ready RL Environments + Evals

Open-source implementation of AlphaEvolve

JDM Editor is an open-source React component for crafting and designing JDM (JSON Decision model) files.

Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities

Implementation and evaluation of the AXIOM architecture from the preprint "AXIOM: Learning to Play Games in Minutes with Expanding Object Centric Models"

Open-source framework for the research and development of foundation models.

📚LeetCUDA: Modern CUDA Learn Notes with PyTorch for Beginners🐑, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.🎉

🏛️A research-friendly codebase for fast experimentation of single-agent reinforcement learning in JAX • End-to-End JAX RL

🪐 The Sebulba architecture to scale reinforcement learning on Cloud TPUs in JAX

Modular, scalable library to train ML models

Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors

Qlib is an AI-oriented Quant investment platform that aims to use AI tech to empower Quant Research, from exploring ideas to implementing productions. Qlib supports diverse ML modeling paradigms, i…

Minimal yet performant LLM examples in pure JAX

React UI + elegant infrastructure for AI Copilots, AI chatbots, and in-app AI agents. The Agentic last-mile 🪁

A Model Context Protocol (MCP) server that enables AI assistants to interact with Kubernetes clusters. It serves as a bridge between AI tools (like Claude, Cursor, and GitHub Copilot) and Kubernetes

A Model Context Protocol (MCP) server that enables AI assistants to interact with Kubernetes clusters. It serves as a bridge between AI tools (like Claude, Cursor, and GitHub Copilot) and Kubernete…

Home for "How To Scale Your Model", a short blog-style textbook about scaling LLMs on TPUs

RADLADS training code

Open-source Windows and Office activator featuring HWID, Ohook, TSforge, KMS38, and Online KMS activation methods, along with advanced troubleshooting.

Timely detections for more proactive and effective actions in offshore oil wells!

C++-based high-performance parallel environment execution engine (vectorized env) for general RL environments.

A Full Live-Scripted CAD Kernel in the Browser

maximal update parametrization (µP)