Boosting 4-bit inference kernels with 2:4 Sparsity

FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.

CUDA on non-NVIDIA GPUs

A general-purpose CV-based framework for extracting precise subtitle timelines from videos with embedded subtitles, from video to .ass file.

🍒 Cherry Studio is a desktop client that supports for multiple LLM providers. Support deepseek-r1

LLM inference in C/C++

An Xposed module to intercept applist detections

The Magic Mask for Android

A Magic Mask to Alter Android System Systemless-ly

Automatic screenshots for Magisk on Android

DRT-o1: Optimized Deep Reasoning Translation via Long Chain-of-Thought

Open-source calculator for LLM system requirements.

OpenAI 接口管理 & 分发系统，改自songquanpeng/one-api。支持更多模型，加入统计页面，完善非openai模型的函数调用。

一个可扩展的通用型小说下载器。

Epson ESC/P-R Driver (generic driver)

宇树科技 Yushu Technology (Unitree) go1 development notes

A userscript for downloading artworks from Pixiv and other websites.

A Flexible Framework for Experiencing Cutting-edge LLM Inference Optimizations

Adds 3D acceleration support for P106-090 / P106-100 / P104-100 / P104-101 / P102-100 / CMP 30HX / CMP 40HX / CMP 50HX / CMP 70HX / CMP 90HX / CMP 170HX mining cards as well as RTX 3060 3840SP, RTX…

📝A simple and elegant markdown editor, available for Linux, macOS and Windows.

An app that brings language models directly to your phone.

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

NVIDIA Linux open GPU with P2P support

OCR离线图片文字识别命令行windows程序，以JSON字符串形式输出结果，方便别的程序调用。提供各种语言API。由 PaddleOCR C++ 编译。

AutoDL平台服务器适配梯子， 使用 Clash 作为代理工具

Yuan 2.0 Large Language Model

🚀 Next Generation AI One-Stop Internationalization Solution. 🚀 下一代 AI 一站式 B/C 端解决方案，支持 OpenAI，Midjourney，Claude，讯飞星火，Stable Diffusion，DALL·E，ChatGLM，通义千问，腾讯混元，360 智脑，百川 AI，火山方舟，新必应，Gemini，Moonshot …

Open source eGPU dock for ROG Ally and ROG Flow