Brandon Ye

hippo-memory. Biologically-inspired memory for AI agents. Decay, retrieval strengthening, consolidation. Zero dependencies.

agency-agents-zh. 🎭 267 个即插即用的 AI 专家角色 — 支持 Hermes Agent/Claude Code/Cursor/Copilot 等 18 种工具，覆盖工程/设计/营销/金融等 20 个部门。含 52 个中国市场原创智能体（小红书/抖音/微信/飞书/钉钉等）。搭配编排器 agency-orchestrator，一句话即可让多位专家按 DAG 自动协作。

phyai. PhyAI is a high-performance framework for running Physical AI models (VLA, WAM, and beyond), supporting both cloud-based serving and on-device deployment.

ai-agent-book. 《深入理解 AI Agent：设计原理与工程实践》（李博杰 著）开源主仓库：全书正文、编译版 PDF 与按章配套代码

homebot. Talk to your home. A voice-driven personal AI agent that chats, assists, and orchestrates your smart home.

open-connector. Open-source auth gateway connecting 1000+ SaaS providers to AI agents through SDK, CLI, MCP, HTTP, and OpenAPI.

qqmusic-skills.

agent-camp. AI Agent 面试知识库：覆盖 LLM、Prompt、RAG、MCP、Tool Use、Agent 架构、Multi-Agent、LangGraph、Claude Code、Codex CLI、工程化评估、安全与源码解析。

agent-skills. Production-grade engineering skills for AI coding agents.

SYSU_Notebook. 本项目分享了中山大学计算机学院本科和研究生阶段的课程资料、笔记、期末考试卷和其他实用的相关资源。希望对同学们的学习有所帮助❤️，如果喜欢记得给个star🌟

OpenCLI. Make Any Website into CLI & Use your logged-in browser by AI agent.

awesome-deepseek-agent.

html-ppt-skill. HTML PPT Studio — AgentSkill with 24 themes, 31 layouts, 20+ animations for building professional HTML presentations

CLI-Anything. "CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/

colleague-skill. 将冰冷的离别化为温暖的 Skill，欢迎加入数字生命1.0！Transforming cold farewells into warm skills? It's giving rebirth era. Welcome to Digital Life 1.0. 🫶

learn-nanobot. 面向小白的 HKUDS/nanobot 面试学习指南 | 17章深度教程 | 134道八股文 | 哆啦A梦漫画图解 | STAR面试法 | 简历模板

nanobot. Ultra-lightweight, open-source, self-hosted personal AI agent framework in Python with WebUI, tools, memory, MCP, multi-agent workflows, automation, and chat apps

Hermes-Wiki. Hermes agent + LLM Wiki + 源代码 完成 Hermes agent wiki

awesome-openclaw-skills. The awesome collection of OpenClaw skills. 5,400+ skills filtered and categorized from the official OpenClaw Skills Registry.🦞

andrej-karpathy-skills. A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

silero-vad. Silero VAD: pre-trained enterprise-grade Voice Activity Detector

unison. Unison file synchronizer

sherpa-onnx. Speech-to-text, text-to-speech, speaker diarization, speech enhancement, source separation, and VAD using next-gen Kaldi with onnxruntime without Internet connection. Support embedded systems, Android, iOS, HarmonyOS, Raspberry Pi, RISC-V, RK NPU, Axera NPU, Ascend NPU, x86_64 servers, websocket server/client, support 12 programming languages

SkillClaw. Let Skills Evolve Collectively with Agentic Evolver

claude-code-best-practice. from vibe coding to agentic engineering - practice makes claude perfect

codex. Lightweight coding agent that runs in your terminal

autoresearch. AI agents running research on single-GPU nanochat training automatically

cli. The official Lark/飞书 CLI tool, maintained by the larksuite team — built for humans and AI Agents. Covers core business domains including Messenger, Docs, Base, Sheets, Calendar, Mail, Tasks, Meetings, and more, with 200+ commands and 20+ AI Agent Skills.

feishu-cli. feishu-cli 是一个功能完整的飞书开放平台命令行工具。它将飞书文档、知识库、电子表格、消息、日历、任务等操作封装为简洁的命令行接口，核心能力是 Markdown ↔ 飞书文档双向无损转换。

metabot. 构建受监督的、自我进化的 Agent 组织的基础设施 | Infrastructure for supervised, self-improving agent organization. 飞书/Telegram 手机端运行 Claude Code 或 Kimi Code（双引擎，两家原生订阅直接用），共享记忆、Agent 工厂、定时任务、通信总线。

PDFMathTranslate. [EMNLP 2025 Demo] PDF scientific paper translation with preserved formats - 基于 AI 完整保留排版的 PDF 文档全文双语翻译，支持 Google/DeepL/Ollama/OpenAI 等服务，提供 CLI/GUI/MCP/Docker/Zotero

claw-code. An agent-managed museum exhibit, built in Rust with Gajae-Code / LazyCodex — developed and maintained with no human intervention.

claude-code-skills. 18 standalone skills for Claude Code and Codex: review, audit, optimization, testing, product discovery, and safe repository publishing.

awesome-claude-skills. A curated list of awesome Claude Skills, resources, and tools for customizing Claude AI workflows

AI-workflow.

nv-sflow. A Python CLI workflow orchestrator with pluggable backends (e.g. local, Slurm) for running declarative YAML DAGs, collecting logs, and organizing outputs consistently.

next-ai-draw-io. A next.js web application that integrates AI capabilities with draw.io diagrams. This app allows you to create, modify, and enhance diagrams through natural language commands and AI-assisted visualization.

Agent-Memory-Paper-List. The paper list of "Memory in the Age of AI Agents: A Survey"

claude-code. Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

PageIndex. 📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG

ECC. The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

openclaw. Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

marker. Convert PDF to markdown + JSON quickly with high accuracy

mac-precision-touchpad. Windows Precision Touchpad Driver Implementation for Apple MacBook / Magic Trackpad

gov-doc-formatter. 基于 LLM agent 的党政机关公文自动排版工具

mini-sglang. A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.

Qingjiao. 面向高校青年教师的纵向科研项目资料汇总仓库

ChinaTextbook. 所有小初高、大学PDF教材。

AISystem. AISystem 主要是指AI系统，包括AI芯片、AI编译器、AI推理和训练框架等AI全栈底层技术

TileRT. Tile-Based Runtime for Ultra-Low-Latency LLM Inference

AI-Research-SKILLs. Comprehensive open-source library of AI research and engineering skills for any AI model. Package the skills and your claude code/codex/gemini agent will be an AI research agent with full horsepower. Maintained by Orchestra Research.

megatron-system-analysis.

Slides-PaperSharing. I am updating my paper-sharing videos on RedNote (小红书). In these videos, I select interesting and important papers that I have read and briefly introduce their most significant aspects in 5 to 8 minutes. I would update my slides for the videos here.

Jupiter. Jupiter is a fast, scalable, and resource-efficient collaborative edge-AI system for generative LLM inference.

awesome-llm-agents. A curated list of awesome LLM agents frameworks.

LightCompress. [EMNLP 2024 & AAAI 2026] A powerful toolkit for compressing large models including LLMs, VLMs, and video generative models.

LightX2V. Lightweight Image Video Action Generation Inference Framework

MineContext. MineContext is your proactive context-aware AI partner（Context-Engineering+ChatGPT Pulse）

leetgpu-challenges. LeetGPU Challenges

Awesome-Multimodal-Large-Language-Models. Reading notes about Multimodal Large Language Models, Large Language Models, and Diffusion Models

CL4R1T4S. LEAKED SYSTEM PROMPTS FOR CHATGPT, CLAUDE, GEMINI, GROK, PERPLEXITY, CURSOR, LOVABLE, REPLIT, AND MORE! - AI SYSTEMS TRANSPARENCY FOR ALL! 👐

vllm-ascend. Community maintained hardware plugin for vLLM on Ascend

pplx-kernels. Perplexity GPU Kernels

rllm. Democratizing Reinforcement Learning for LLMs

Paper2Poster. [NeurIPS 2025] Open-source Multi-agent Poster Generation from Papers

cupy. NumPy & SciPy for GPU

DeepSeek-V3. Python

MegaScale-Infer-Prototyp. Prototyp MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism

nano-vllm. Nano vLLM

ml-fastvlm. This repository contains the official implementation of "FastVLM: Efficient Vision Encoding for Vision Language Models" - CVPR 2025

cuda-python. CUDA Python: Performance meets Productivity

StreamMind. [ICCV 2025] StreamMind: Unlocking Full Frame Rate Streaming Video Dialogue through Event-Gated Cognition

awesome-mcp-servers. A collection of MCP servers.

LLaVA-NeXT. Python

CUDA-Learn-Notes. 📚200+ Tensor/CUDA Cores Kernels, ⚡️flash-attn-mma, ⚡️hgemm with WMMA, MMA and CuTe (98%~100% TFLOPS of cuBLAS/FA2 🎉🎉).

lite.ai.toolkit. 🛠 A lite C++ toolkit: contains 100+ Awesome AI models, support MNN, NCNN, TNN, ONNXRuntime and TensorRT. 🎉🎉

ffpa-attn. FFPA: Kernel Library for Large Headdim Attention - 1.5x~6x speedup over PyTorch SDPA.

lite.ai.toolkit. A lite C++ AI toolkit: 100+ models with MNN, ORT and TRT, including Det, Seg, Stable-Diffusion, Face-Fusion.

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

open-r1. Fully open reproduction of DeepSeek-R1

open-infra-index. Production-tested AI infrastructure tools for efficient AGI development and community-driven innovation

Torch-Pruning. [CVPR 2023] DepGraph: Towards Any Structural Pruning; LLMs, Vision Foundation Models, etc.

Awesome-ML-SYS-Tutorial. My learning notes for ML SYS.

Awesome-KV-Cache-Compression. 📰 Must-read papers on KV Cache Compression (constantly updating 🤗).

exo. Run frontier AI locally.

Thinking-Claude. Let your Claude able to think

lingua. Meta Lingua: a lean, efficient, and easy-to-hack codebase to research LLMs.

O1-Journey. O1 Replication Journey

SLM_Survey.

VITA. ✨✨[NeurIPS 2025] VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction

mi-gpt. 🏠 将小爱音箱接入 ChatGPT 和豆包，改造成你的专属语音助手。

xDiT. xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism

llm_interview_note. 主要记录大语言大模型（LLMs） 算法（应用）工程师相关的知识及面试题

simple-computer. the scott CPU from "But How Do It Know?" by J. Clark Scott

vllmini. A minimal implementation of vllm.

mem0. Universal memory layer for AI Agents

ring-attention-pytorch. Implementation of 💍 Ring Attention, from Liu et al. at Berkeley AI, in Pytorch

bitsandbytes. Accessible large language models via k-bit quantization for PyTorch.

lectures. Material for gpu-mode lectures

Mooncake. Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.

SpeculativeDecodingPapers. 📰 Must-read papers and blogs on Speculative Decoding ⚡️

swiftLLM. A tiny yet powerful LLM inference system tailored for researching purpose. vLLM-equivalent performance with only 2k lines of code (2% of vLLM).

llama-sandbox. A collection of experiments related to LLM inference with llama.cpp/mlx

xiaogpt. Play ChatGPT and other LLM with Xiaomi AI Speaker

DeepSeek-V2. DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model

Awesome-Chinese-LLM. 整理开源的中文大语言模型，以规模较小、可私有化部署、训练成本较低的模型为主，包括底座模型，垂直领域微调及应用，数据集与教程等。

llamafile. Distribute and run LLMs with a single file.

llm.c. LLM training in simple, raw C/CUDA

generative-ai-for-beginners. 21 Lessons, Get Started Building with Generative AI

scattermoe. Triton-based implementation of Sparse Mixture of Experts.

Awesome-Parameter-Efficient-Transfer-Learning. Collection of awesome parameter-efficient fine-tuning resources.

Galaxy-MNN. Work in progress project upon Alibaba MNN.

LLMSys-PaperList. Large Language Model (LLM) Systems Paper List

Awesome-LLM-System-Papers.

VideoSys. VideoSys: An easy and efficient system for video generation

awesome-large-multimodal-agents.

AgentBench. A Comprehensive Benchmark to Evaluate LLMs as Agents (ICLR'24)

bert4torch. An elegent pytorch implement of transformers

LWM. Large World Model -- Modeling Text and Video with Millions Context

llama-cpp-wasm. WebAssembly (Wasm) Build and Bindings for llama.cpp

everyone-can-use-english. 人人都能用英语

mixtral-offloading. Run Mixtral-8x7B models in Colab or consumer desktops

mllm. Fast Multimodal LLM on Mobile Devices

distributed-llama. Distributed LLM inference. Connect home devices into a powerful cluster to accelerate LLM inference. More devices means faster inference.

x-stable-diffusion. Real-time inference for Stable Diffusion - 0.88s latency. Covers AITemplate, nvFuser, TensorRT, FlashAttention. Join our Discord communty: https://discord.com/invite/TgHXuSJEk6

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

chatllm.cpp. Pure C++ implementation of several models for real-time chatting on your computer (CPU & GPU)

Awesome_Multimodel_LLM. Awesome_Multimodel is a curated GitHub repository that provides a comprehensive collection of resources for Multimodal Large Language Models (MLLM). It covers datasets, tuning techniques, in-context learning, visual reasoning, foundational models, and more. Stay updated with the latest advancement.

Personal_LLM_Agents_Survey. Paper list for Personal LLM Agents

SwiftInfer. Efficient AI Inference & Serving

Cook_Notebook. 打工人做饭日记

llama-moe. ⛷️ LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-training (EMNLP 2024)

mali. Everything we learnt from hacking Arm Mali GPUs.

gpt-fast. Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

llamafia.github. Python

LookaheadDecoding. [ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decoding

gloo. Collective communications library with various primitives for multi-machine training.

fastmoe. A fast MoE impl for PyTorch

system-design-101. Explain complex systems using visuals and simple terms. Help you prepare for system design interviews.

gpu_poor. Calculate token/s & GPU memory requirement for any LLM. Supports llama.cpp/ggml/bnb/QLoRA quantization

Generalizable-Mixture-of-Experts. GMoE could be the next backbone model for many kinds of generalization task.

TensorRT-LLM. TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way.

LLMsPracticalGuide. A curated list of practical guide resources of LLMs (LLMs Tree, Examples, Papers)

Awesome-Efficient-LLM. A curated list for Efficient Large Language Models

Medusa. Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads

calculon. Python

LLMSpeculativeSampling. Fast inference from large lauguage models via speculative decoding

llama-gpt. A self-hosted, offline, ChatGPT-like chatbot. Powered by Llama 2. 100% private, with no data leaving your device. New: Code Llama support!

awesome-llm-and-aigc. 🚀🚀🚀A collection of some awesome public projects about Large Language Model(LLM), Vision Language Model(VLM), Vision Language Action(VLA), AI Generated Content(AIGC), the related Datasets and Applications.

LightLLM. LightLLM is a Python-based LLM (Large Language Model) inference and serving framework, notable for its lightweight design, easy scalability, and high-speed performance.

OpenMoE. A family of open-sourced Mixture-of-Experts (MoE) Large Language Models

llvmlite. A lightweight LLVM python binding for writing JIT compilers

Awesome-LLM. Awesome-LLM: a curated list of Large Language Model

llama2.c. Inference Llama 2 in one file of pure C

TransformerEngine. A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.

astra-sim. ASTRA-sim2.0: Modeling Hierarchical Networks and Disaggregated Systems for Large-model Training at Scale

tensor_parallel. Automatically split your PyTorch models on multiple GPUs for training & inference

annotated_deep_learning_paper_implementations. 🧑‍🏫 60+ Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers (original, xl, switch, feedback, vit, ...), optimizers (adam, adabelief, sophia, ...), gans(cyclegan, stylegan2, ...), 🎮 reinforcement learning (ppo, dqn), capsnet, distillation, ... 🧠

LLMSurvey. The official GitHub page for the survey paper "A Survey of Large Language Models".

chatGLM-6B-QLoRA. 使用peft库，对chatGLM-6B/chatGLM2-6B实现4bit的QLoRA高效微调，并做lora model和base model的merge及4bit的量化（quantize）。

nlp_chinese_corpus. 大规模中文自然语言处理语料 Large Scale Chinese Corpus for NLP

fastllm. fastllm是后端无依赖的高性能大模型推理库。同时支持张量并行推理稠密模型和混合模式推理MOE模型，任意10G以上显卡即可推理满血DeepSeek。双路9004/9005服务器+单显卡部署DeepSeek满血满精度原版模型，单并发20tps；INT4量化模型单并发30tps，多并发可达60+。

flash-attention. Fast and memory-efficient exact attention

InferLLM. a lightweight LLM model inference framework

The-Art-of-Linear-Algebra. Graphic notes on Gilbert Strang's "Linear Algebra for Everyone"

fairseq. Facebook AI Research Sequence-to-Sequence Toolkit written in Python.

whisper. Robust Speech Recognition via Large-Scale Weak Supervision

d2go. D2Go is a toolkit for efficient deep learning

MNN-APPLICATIONS. MNN applications by MNN, JNI exec, RK3399. Support tflite\tensorflow\caffe\onnx models.

vllm. A high-throughput and memory-efficient inference and serving engine for LLMs

tutorials. This repository contains tutorials and examples for Triton Inference Server

VisualGLM-6B. Chinese and English multimodal conversational language model | 多模态中英双语对话语言模型

AI-Papers-of-the-Week. 🔥Highlighting the top ML papers every week.

gpt-2-Pytorch. Simple Text-Generator with OpenAI gpt-2 Pytorch Implementation

private-gpt. Complete API layer for private AI applications on local models: RAG, skills, tools, MCP, text-to-sql, and more. Works with any OpenAI-compatible inference server.

open-llms. 📋 A list of open LLMs available for commercial use.

logseq. A privacy-first, open-source platform for knowledge management and collaboration. Download link: http://github.com/logseq/logseq/releases. roadmap: https://logseq.io/p/NX4mc_ggEV

open_llama. OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset

mlc-llm. Universal LLM Deployment Engine with ML Compilation

libai. LiBai(李白): A Toolbox for Large-Scale Distributed Parallel Training

Raspberry_torch. Self-compiled Torch wheels(torch 1.8 & torch 1.6) for the Raspberry PI ARMV7(Raspberry 4b(+)).

mnn-llm. llm deploy project based mnn. This project has merged into MNN.

ChatGLM-6B. ChatGLM-6B: An Open Bilingual Dialogue Language Model | 开源双语对话语言模型

GPTCache. Semantic cache for LLMs. Fully integrated with LangChain and llama_index.

vicuna-7b. Vicuna 7B is a large language model that runs in the browser. Exposes programmatic access with minimal configuration.

MOSS. An open-source tool-augmented conversational language model from Fudan University

llama. Inference code for Llama models

self-instruct. Aligning pretrained language models with instruction data generated by themselves.

wasm-gpt. Tensor library for machine learning

ggml. Tensor library for machine learning

FedModule. 联邦学习模块化框架，支持各类FL。A universal federated learning framework, free to switch thread and process modes

MiniGPT-4. Open-sourced codes for MiniGPT-4 and MiniGPT-v2 (https://minigpt-4.github.io, https://minigpt-v2.github.io/)

awesome-AI-system. paper and its code for AI System

web-llm. High-performance In-browser LLM Inference Engine

zotero-gpt. GPT Meet Zotero.

AutoGPT. AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.

baize-chatbot. Let ChatGPT teach your own chatbot in hours with a single GPU!

openplayground. An LLM playground you can run on your laptop

FastChat. An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.

server. The Triton Inference Server provides an optimized cloud and edge inferencing solution.