Out-of-tree Linux thunderbolt_net driver (DKMS) tracking upstream, with a receive-side workaround for macOS Thunderbolt Bridge TCP aggregates: validated GSO reconstruction bounded by each flow's le…

Dogfooding vLLM backport for older GPUs.

A local inference engine for Apple silicon, built around the model.

Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported

Open-source CUDA, Triton and HIP compiler targeting multiple GPU and CPU architectures.

A unified library for building, evaluating, and storing speculative decoding algorithms for LLM inference in vLLM

Virtualized Elastic KV Cache for Dynamic GPU Sharing and Beyond

A machine learning compiler for GPUs, CPUs, and ML accelerators

Low-latency collective transport and distributed LLM serving stack for switchless GB10 clusters.

Run iOS/ macOS in Qemu. Virtual iPhone 17, 16, 15, 14, 13, 12 and M5-M1 Apple Si Macs supported.

OMY (Open Model Yard) is created to ensure the continued development of the Transformers project adheres to a true open source movement — open to all architectures, platforms, and both public and p…

Training neural networks on Apple Neural Engine via reverse-engineered private APIs

1 script run the vphone

A high-performance inference engine for AI models

Pure-Python Git implementation

Reprodcution/Validation of Recirculation paper

ShadowVPN: UDP PSK user-mode VPN with shadowsocks AEAD

Live GPU emulator for High-Bandwidth Flash: applies HBF timing, capacity and thermal effects to a real LLM inference workload while it executes on a real GPU.

DeepSeek Harness: Everything is a Plugin.

Large-scale LLM inference engine

TokenSpeed is a speed-of-light LLM inference engine.

Optimized local serving engine for Kimi-Linear-48B: INT4 quantizer, fused decode kernels for a measured 3.18x, and an OpenAI-compatible server. Ships with k3, a bridge that detects the client per r…

Lint modeling, modular, and configuration files for structural conventions

13 keyboard with 3d

A PyTorch-based framework for Quantum Classical Simulation, Quantum Machine Learning, Quantum Neural Networks, Parameterized Quantum Circuits with support for easy deployments on real quantum compu…

My reasearch of losslessly compressing LLM weights.

Google's Engineering Practices documentation