Download WireGuard for Windows at https://www.wireguard.com/install . This repo is a mirror only. Official repository is at https://git.zx2c4.com/wireguard-windows

A simple, decentralized mesh VPN with WireGuard support.

Turn Antigravity / codex / github copilot into Anthropic & Openai API compatible server. Usable with Claude Code / Xcode etc.

✨ Reverse-engineered Python API for Google Gemini web app

A service that can convert ChatGPT on the web to OpenAI API format.

A computational fluid dynamics solver under development for use in computer programming and fluid mechanics practical exercises

55nm CMOS Open Source PDK by ICsprout Integrated Circuit Co., Ltd.

Foundation Models for Time Series

Open-source CUDA compiler targeting multiple GPU architectures. Compiles .cu to AMD and Tenstorrent GPU's

LiveUSB Bootable exploit chain to unlock all features of xx30 ThinkPad machines. WiFi Whitelist, Advanced Menu, Overclocking.

A fast implementation of T5/UL2 in PyTorch using Flash Attention

Super fast FP32 matrix multiplication on RDNA3

Mozilla for Mac OS X Tiger and Leopard on PowerPC.

Acceleration codes for the Ozaki-scheme on integer matrix multiplication units.

PDLP algorithm for linear programming

Yet Another ARM SVE Emulator -- A simple trap hander of illegal instructions on aarch64

OLMoE: Open Mixture-of-Experts Language Models

MS-DOS compilers, assemblers, and interpreters

Accelerating Physics Informed Neural Networks (PINNs) using Meshless Discretizations

Fault tolerance for PyTorch (HSDP, LocalSGD, DiLoCo, Streaming DiLoCo)

Official Github repository for the SIGCOMM '24 paper "Accelerating Model Training in Multi-cluster Environments with Consumer-grade GPUs"

a simple Flash Attention v2 implementation with ROCM (RDNA3 GPU, roc wmma), mainly used for stable diffusion(ComfyUI) in Windows ZLUDA environments.

Fast and memory-efficient exact attention

A simple, performant and scalable Jax LLM!

JAX + Tinygrad = Run everywhere

AI Tensor Engine for ROCm

A hackable library for running and fine-tuning modern transformer models on commodity and alternative GPUs, powered by tinygrad.