- pytorch by pytorch, 106k stars. Python deep learning framework with tensors, autograd and GPU acceleration
- vllm by vllm-project, 97k stars. High-throughput inference and serving engine for large language models
- ghostty by ghostty-org, 64k stars. GPU-accelerated terminal emulator using native UI on each platform
- voicebox by jamiepine, 61k stars. Voice studio for cloning voices, dictation and generating speech
- AnyPS5 by boykopovar, 54k stars. Tool that automatically ports PS5 executables to run on Linux and Windows
- colibri by JustVugg, 51k stars. Inference engine for MoE models that spreads experts across GPU, RAM and NVMe
- DeepSpeed by deepspeedai, 43k stars. Library for distributed training and inference of large PyTorch models
- airllm by lyogavin, 42k stars. Runs large language models layer by layer so they fit in a small GPU's memory
- ds4 by antirez, 24k stars. Local inference engine for DeepSeek models on Metal, CUDA and ROCm
- gpu-unlock by Zagitta, 20k stars. Tool that claims to lift FP16 and TF32 tensor throughput on GeForce 20 to 40 series
- Speech by NVIDIA-NeMo, 19k stars. Framework for speech recognition, text-to-speech, and translation models
- Waifu2x-Extension-GUI by AaronFeng753, 17k stars. Windows GUI for upscaling images, GIFs and video with several AI models
- TensorRT-LLM by NVIDIA, 15k stars. Python API and runtime for fast LLM inference on NVIDIA GPUs
- cupy by cupy, 12k stars. NumPy and SciPy compatible array library that runs on NVIDIA GPUs
- renderdoc by baldurk, 11k stars. Standalone graphics debugger for capturing and inspecting frames in Vulkan and D3D
- KytyPS5 by KytyPS5, 11k stars. Translation layer for running PlayStation 5 code on Windows, Linux and macOS
- horizon by horizon-rust, 9.8k stars. Minecraft Java Edition reimplemented in Rust, with a launcher
- DeepGEMM by deepseek-ai, 9.2k stars. GPU matrix multiplication kernel library written in CUDA
- tilelang by tile-ai, 9.0k stars. Python DSL for writing high-performance GPU and CPU kernels
- modly by lightningpixel, 8.7k stars. Desktop app that makes 3D models from images or text prompts on your GPU
- ODS by Osmantic, 7.8k stars. Pre-release stack that turns a PC or Mac into a private AI server with Docker
- magnitude by magnitudedev, 7.8k stars. Inference engine that tunes kernels per hardware for faster model runs
- turbo-fieldfare by drumih, 7.4k stars. Runs a Gemma model on Apple Silicon Macs with low memory use via Metal
- openscreen by getopenscreen, 7.3k stars. Cross-platform screen recorder for product demos, built on Electron
- rune by anomalyco, 7.2k stars. GPU-rendered native UI framework in Rust for desktop and the web
- koharu by koharu-rs, 6.7k stars. Machine learning tool that translates manga pages, written in Rust
- Model-Optimizer by NVIDIA, 5.6k stars. Library for quantization, pruning, distillation and speculative decoding
- gpu-perf-engineering-resources by wafer-ai, 4.4k stars. Curated reading list on GPU fundamentals and production inference performance
- Mac.apk by kksimp, 4.2k stars. Runs Android apps natively on Apple Silicon Macs
- ramalama by containers, 3.1k stars. CLI that serves AI models locally in containers for inference
- openlake by openlake-project, 3.1k stars. High-throughput storage engine aimed at LLM inference and GPU training
- lucebox by Luce-Org, 3.0k stars. Inference server that speeds up local LLMs on consumer GPUs
- hakios-xnu-kernel by oosexgang, 2.8k stars. Experimental port of Apple's XNU kernel that boots on Android phones from RAM
- manimgx by academa-labs, 2.7k stars. Manim-compatible Python API for fast math and 3D animations on a GPU core
- HyperQwen by syv-ai, 2.7k stars. vLLM setup and quantization tuning for serving Qwen models on a single GPU
- vgpu by vercel-labs, 2.7k stars. Cross-runtime WebGPU library for shaders, 3D scenes and GPU tensors
- BUI by ppmpreetham, 2.6k stars. Blender's GPU-accelerated UI toolkit extracted for other languages
- BRUTAL by ahwootech, 2.5k stars. C# framework for games and game engines with direct Vulkan access
- spirula-studio by harry7557558, 2.4k stars. Gaussian splat trainer that turns video into splats and meshes on Vulkan or CUDA
- YuE2-Studio by timoncool, 2.2k stars. Native Windows app that generates full songs locally with an editable score
New repos every two days, stars refreshed daily. Visits are counted without cookies, and the numbers are public. How it works Data RSS Newsletter