Ma Jimmy

lingbot-vision. Self-supervised learning for spatial perception

InternNav. InternRobotics' open platform for building generalized navigation foundation models.

Eagle. Eagle: Frontier Vision-Language Models with Data-Centric Strategies

WorldDirector. Python

SM3Det. [AAAI26 Oral] "SM3Det: A Unified Model for Multi-Modal Remote Sensing Object Detection" and [ICML 26] "Unifying Heterogeneous Multi-Modal Remote Sensing Detection Via Language-Pivoted Pretraining"

Echo-Memory. A Simple Baseline for Video World Models with Memory

tiger. Open Source LLM toolkit to build trustworthy LLM applications. TigerArmor (AI safety), TigerRAG (embedding, RAG), TigerTune (fine-tuning)

cosmos-cookbook. Post-training scripts and samples for NVIDIA Cosmos ecosystem

sapiens2. 1K resolution vision transformers pretrained on 1B human images.

agent-network. 助力搭建你的数字 AI 员工军团 — 多 Agent 一行命令组网协作。Claude Code / Claude Agent SDK / Codex / Grok Build 4 runtime + 8+ 家 LLM（Anthropic / OpenAI / xAI / MiniMax / DeepSeek / GLM / Kimi / 书生 / 小米 MiMo），自带 Web Dashboard，Apache 2.0 开源。

learning-beyond-gradients. Heuristic Learning Blog Post

ai-model-workbench. Obsidian desktop plugin: Babylon.js 3D model viewer with multi-model grid, presets, knowledge notes. Supports GLB/GLTF/STL/OBJ/SPLAT.

Fraud-R1. [ACL 2025 Findings] Fraud-R1 : A Multi-Round Benchmark for Assessing the Robustness of LLM Against Augmented Fraud and Phishing Inducements

OmniVoice. High-Quality Voice Cloning TTS for 600+ Languages

Claw-R1. Claw-R1: Empowering OpenClaw with Advanced Agentic RL.

SmsForwarder. 短信转发器——监控Android手机短信、来电、APP通知，并根据指定规则转发到其他手机：钉钉群自定义机器人、钉钉企业内机器人、企业微信群机器人、飞书机器人、企业微信应用消息、邮箱、bark、webhook、Telegram机器人、Server酱、PushPlus、手机短信等。包括主动控制服务端与客户端，让你轻松远程发短信、查短信、查通话、查话簿、查电量等。（V3.0 新增）PS.这个APK主要是学习与自用，如有BUG请提ISSUE，同时欢迎大家提PR指正

I_am_a_person. 实时互动的GPT数字人技术笔记，不是数字人开源项目

mcore-bridge. MCore-Bridge: Providing Megatron-Core model definitions for state-of-the-art large language models and making Megatron training as simple as Transformers.

mcore-bridge. MCore-Bridge: Providing Megatron-Core model definitions for state-of-the-art large models and making Megatron training as simple as Transformers — with support for 300+ large language models (Qwen3-Next, GLM-5.1, Deepseek-V4, MiniMax-2.7, ...) and 200+ multimodal large models (Qwen3.5, Qwen3-Omni, Gemma4, ...).

claude-code-rev. Runnable ClaudeCode source code

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

opencode. The open source coding agent.

skills. C#

TreeSkill. Skill optimization framework for LLMs — evolve system prompts via textual gradient descent with beam search, human-in-the-loop annotation, and DPO preference data export.

EvoSkill. Skill optimization framework for LLMs — evolve system prompts via textual gradient descent with beam search, human-in-the-loop annotation, and DPO preference data export.

nanoclaw. A lightweight alternative to OpenClaw that runs in containers for security. Connects to WhatsApp, Telegram, Slack, Discord, Gmail and other messaging apps,, has memory, scheduled jobs, and runs directly on Anthropic's Agents SDK

Dasheng-Audiogen. Generate a complete audio clip with music, intelligible speech, and sound effects from text in one pass.

self-evolving-agent. An OpenClaw skill that upgrades self-improving agents from reactive error logging to goal-driven capability evolution with curriculum, evaluation, transfer, and promotion.

OpenClaw-RL. OpenClaw-RL: Train any agent simply by talking

twinkle. Twinkle✨: Training workbench to make your model glow.

transformers. 🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

ROLL. An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models

moonshine. Very low latency speech to text, intent recognition, and text to speech, for building voice agents and interfaces

Multi-Turn-RL-Agent. Python

RSCoVLM. [Remote Sensing 2026] Co-Training Vision Language Models for Remote Sensing Multi-task Learning

FinRL. FinRL®: Financial Reinforcement Learning. 🔥

Odyssey-Arena. Extremely Long-Horizon Agentic Tasks Requiring Active Acting and Inductive Reasoning

Self-Aligned-Reward-Towards_Effective_and_Efficient_Reasoners. Python

vllm-omni. A framework for efficient model inference with omni-modality models

llmscope. Python

LLM-Powered-Life-Remake. LLM-Powered Life Restart Simulator

ManiSkill. Manipulation Skill Framework, an open source GPU parallelized robotics simulator and benchmark

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

RoboCerebra. https://arxiv.org/pdf/2506.06677

Open-AutoGLM. An Open Phone Agent Model & Framework. Unlocking the AI Phone for Everyone

VideoMind. VideoMind: An Omni-Modal Video Dataset with Intent Grounding for Deep-Cognitive Video Understanding

DeepEyesV2. Python

reader3. Quick illustration of how one can easily read books together with LLMs. It's great and I highly recommend it.

agent-lightning. The absolute trainer to light up AI agents.

HI-TransPA. HI-TransPA: Hearing Impairments Translation Personal Assistant

MiniMind-in-Depth. 轻量级大语言模型MiniMind的源码解读，包含tokenizer、RoPE、MoE、KV Cache、pretraining、SFT、LoRA、DPO等完整流程

slime. slime is an LLM post-training framework for RL Scaling.

skrl. Modular Reinforcement Learning (RL) library (implemented in PyTorch, JAX, and NVIDIA Warp) with support for Gymnasium/Gym, NVIDIA Isaac Lab, MuJoCo Playground and other environments

kernels-community. Kernel sources for https://huggingface.co/kernels-community

PaddleOCR. Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.

MUG-V. MUG-V 10B: High-efficiency Training Pipeline for Large Video Generation Models

OmniVinci. OmniVinci is an omni-modal LLM for joint understanding of vision, audio, and language.

LAE-DINO. [AAAI'25] Official Code for “Locate Anything on Earth: Advancing Open-Vocabulary Object Detection for Remote Sensing Community"

OpenCE. OpenCE (Open Context Engineering): A community toolkit to implement, evaluate, and combine LLM context strategies (RAG, ACE, Compression). Evolved from the `ACE-open` reproduction.

Rex-Omni. [CVPR2026] Detect Anything via Next Point Prediction

nanochat. The best ChatGPT that $100 can buy.

openvla. OpenVLA: An open-source vision-language-action model for robotic manipulation.

openvla-oft. Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success

LLMs-from-scratch. Implement a ChatGPT-like LLM in PyTorch from scratch, step by step

Echo-4o. Echo-4o: Harnessing Proprietary Models’ Synthetic Images for Improved Image Generation

index-tts. An Industrial-Level Controllable and Efficient Zero-Shot Text-To-Speech System

VeOmni. VeOmni: Scaling Any Modality Model Training with Model-Centric Distributed Recipe Zoo

VTubeStudio. VTube Studio API Development Page

Open-LLM-VTuber. Talk to any LLM with hands-free voice interaction, voice interruption, and Live2D taking face running locally across platforms

Qwen3-Omni. Qwen3-omni is a natively end-to-end, omni-modal LLM developed by the Qwen team at Alibaba Cloud, capable of understanding text, audio, images, and video, as well as generating speech in real time.

MS-SNSD. The Microsoft Scalable Noisy Speech Dataset (MS-SNSD) is a noisy speech dataset that can scale to arbitrary sizes depending on the number of speakers, noise types, and Speech to Noise Ratio (SNR) levels desired.

MM-Self-Improve. A Self-Training Framework for Vision-Language Reasoning

AV-HuBERT-S2S. Huggingface Implementation of AV-HuBERT on the MuAViC Dataset

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

VSP-LLM. Python

muavic. MuAViC: A Multilingual Audio-Visual Corpus for Robust Speech Recognition and Robust Speech-to-Text Translation

LLaVA-OneVision-2. Fully Open Framework for Democratized Multimodal Training

InternVLA-M1. InternVLA-M1: A Spatially Guided Vision-Language-Action Framework for Generalist Robot Policy

Awesome-RL-for-LRMs. A Survey of Reinforcement Learning for Large Reasoning Models

MuseTalk. MuseTalk: Real-Time High Quality Lip Synchorization with Latent Space Inpainting

Uni-Sign. [ICLR'25] Official Implement of "Uni-Sign: Toward Unified Sign Language Understanding at Scale"

super-agent-party. ⭐ All-in-one AI companion! Super Agent Party = Self hosted neuro sama + openclaw! ⭐ 全能AI伴侣！超级智能体派对 = 自托管neuro sama + openclaw!

disen-speech-tokenizer. 一个面向解藕语音属性的tokenizer，可以用来解藕语音属性、提升音频分类（如鼾声分类）算法性能

seed-oss. Python

LegalDelta. [ICASSP '26] This is the code repo for our paper: LegalΔ: Enhancing Legal Reasoning in LLMs via Reinforcement Learning with Chain-of-Thought Guided Information Gain.

ResumeSample. Resume template for Chinese programmers . 程序员简历模板系列。包括PHP程序员简历模板、iOS程序员简历模板、Android程序员简历模板、Web前端程序员简历模板、Java程序员简历模板、C/C++程序员简历模板、NodeJS程序员简历模板、架构师简历模板以及通用程序员简历模板

MindOmni. [NeurIPS2025] The official implementation of MindOmni: Unleashing Reasoning Generation in Vision Language Models with RGPO

VLM-formula-recognition-dataset. 书生实战营第六期打榜数据集测评方案

TGRS-HRRSD-Dataset. MATLAB

Vision-Language-Transformer. [ICCV2021 & TPAMI2023] Vision-Language Transformer and Query Generation for Referring Segmentation

Book-Mathematical-Foundation-of-Reinforcement-Learning. This is the homepage of a new book entitled "Mathematical Foundations of Reinforcement Learning."

dasheng-lm. Efficient audio understanding with general audio captions

new-api. A unified AI model hub for aggregation & distribution. It supports cross-converting various LLMs into OpenAI-compatible, Claude-compatible, or Gemini-compatible formats. A centralized gateway for personal and enterprise model management. 🍥

train-higgs-audio. Text-audio foundation model from Boson AI

Qwen3-SmVL. 将SmolVLM2的视觉头与Qwen3-0.6B模型进行了拼接微调

higgs-audio. Text-audio foundation model from Boson AI

OpenJudge. OpenJudge: A Unified Framework for Holistic Evaluation and Quality Rewards

True-Story-of-Pangu. 诺亚盘古大模型研发背后的真正的心酸与黑暗的故事。

BongoCat. 🐱 跨平台互动桌宠 BongoCat，为桌面增添乐趣！

Lip2Speech. A pipeline to read lips and generate speech for the read content, i.e Lip to Speech Synthesis.

VLM-R1. Solve Visual Understanding with Reinforced VLMs

SuperClaude_Framework. A configuration framework that enhances Claude Code with specialized commands, cognitive personas, and development methodologies.

TradingAgents. TradingAgents: Multi-Agents LLM Financial Trading Framework

meanflow. JAX implementation of MeanFlow

nano-vllm. Nano vLLM

3DFace. This repository contains the dataset including the pair of 2D face image and its corresponding 3D face geometry model.

DeepEyes. Python

DeepSpeedExamples. Example models using DeepSpeed

ThinkGeo. [CVPR 2026 🔥] ThinkGeo is a Comprehensive Benchmark to evaluate Tool-Augmented Agents for Remote Sensing Tasks

BLIP3o. Official implementation of BLIP3o-Series

minimind-v. 👀「大模型」2小时从0训练65M参数的视觉多模态VLM！Train a 65M-parameter VLM from scratch in just 2h!

nanoVLM. The simplest, fastest repository for training/finetuning small-sized VLMs.

STAR. [CVPR 2023] STAR Loss: Reducing Semantic Ambiguity in Facial Landmark Detection

Index-anisora. Python

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

Awesome-Autoregressive-Visual-Generation. This is a repo to track the latest autoregressive visual generation papers.

ChinaTextbook. 所有小初高、大学PDF教材。

AlignBench. 大模型多维度中文对齐评测基准 (ACL 2024)

Kimi-Audio. Kimi-Audio, an open-source audio foundation model excelling in audio understanding, generation, and conversation

chinese-llm-benchmark. 非线智能 NoneLinear - ReLE评测：中文AI大模型能力评测（持续更新）：目前已囊括374个大模型，覆盖chatgpt、gpt-5.4、谷歌gemini-3.1-pro、Claude-4.6、文心ERNIE-X1.1、ERNIE-5.0、qwen3.6-max、qwen3.6-plus、百川、讯飞星火、商汤senseChat等商用模型， 以及step3.5-flash、kimi-k2.6、ernie4.5、MiniMax-M2.7、deepseek-v4、Qwen3.6、llama4、智谱GLM-5.1、MiMo-V2、LongCat、gemma4、mistral等开源大模型。不仅提供排行榜，也提供规模超200万的大模型缺陷库！方便广大社区研究分析、改进大模型。

minimind. 🧠「大模型」2小时完全从0训练64M的小参数LLM！Train a 64M-parameter LLM from scratch in just 2h!

align-anything. Align Anything: Training All-modality Model with Feedback

GRPO-Zero. Implementing DeepSeek R1's GRPO algorithm from scratch

GraphGen. GraphGen: Enhancing Supervised Fine-Tuning for LLMs with Knowledge-Driven Synthetic Data Generation

modelscope-classroom. Jupyter Notebook

WeClone. 🚀 One-stop solution for creating your AI twin from chat history 💡 Fine-tune LLMs with your chat logs to capture your unique style, then bind to a chatbot to bring your digital self to life.

SAFE-QAQ. [ACL '26] SAFE-QAQ: End-to-End Slow-Thinking Audio-Text Fraud Detection via Reinforcement Learning.

InternVL-MMDetSeg. Train InternViT-6B in MMSegmentation and MMDetection with DeepSpeed

TeleAntiFraud. [MM '25]TeleAntiFraud-28k, the first open-source audio-text slow-thinking dataset specifically designed for automated telecom fraud analysis.

GLM-4-Voice. GLM-4-Voice | 端到端中英语音对话模型

MMEarth-data. This repository contains code to download data for the preprint "MMEarth: Exploring Multi-Modal Pretext Tasks For Geospatial Representation Learning"

Matcha-TTS. [ICASSP 2024] 🍵 Matcha-TTS: A fast TTS architecture with conditional flow matching

dify-mcp-server. Model Context Protocol (MCP) Server for dify workflows

VisualThinker-R1-Zero. Explore the Multimodal “Aha Moment” on 2B Model

tts-now. 跨平台基于云平台(阿里云、讯飞等)语音合成 API 的文字转语音助手。支持单文本快速合成和批量合成。支持windows、macOS、Linux。

Awesome-Remote-Sensing-Foundation-Models.

OmniAlign-V. Official Repository of ACL 2025 paper OmniAlign-V: Towards Enhanced Alignment of MLLMs with Human Preference

RSAR. [CVPR 2025] Official implementation for the paper "RSAR: Restricted State Angle Resolver and Rotated SAR Benchmark".

ChatTTS-ui. 一个简单的本地网页界面，使用ChatTTS将文字合成为语音，同时支持对外提供API接口。A simple native web interface that uses ChatTTS to synthesize text into speech, along with support for external API interfaces.

talk2ai. Python

MM-RLHF. The Next Step Forward in Multimodal LLM Alignment

aclpubcheck. Tools for checking ACL paper submissions

ms-swift. Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.6, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM4.5v, Gemma4, Llava, Phi4, ...) (AAAI 2025).

LANCE. Official repository of Language Models as Continuous Self-Evolving Data Engineers

LLaVA-CoT. [ICCV 2025] LLaVA-CoT, a visual language model capable of spontaneous, systematic reasoning

Qwen3-VL. Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.

SARChat. The first large-scale multimodal dialogue dataset focusing on Synthetic Aperture Radar (SAR) imagery.

LlamaFactory. Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

verl. verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

ModernBERT. Bringing BERT into modernity via both architecture changes and scaling

argilla-cookbook. Simple examples using Argilla tools to build AI

LLM2CLIP. LLM2CLIP significantly improves already state-of-the-art CLIP models.

genesis-world. Simulation platform for general-purpose robotics & embodied AI learning.

LLM-Ceping. 全方位大模型评测知识库 | 提示词工程（Prompt Engineer）、各渠道大模型榜单（LeaderBoard）、标杆数据集、安全检测、对抗攻击、智能体、优质数据、文本分类、关系抽取、语音识别、语音合成、多模态、文本生成图片、文本生成视频、点云、智能对话、摘要总结、问答、字词释义、知识问答、学科考试、医学、信息学、心理学、指代消解、常识推理、知识问答...

TripletCLIP. [NeurIPS 2024] Official PyTorch implementation of "Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives"

Vitron. NeurIPS 2024 Paper: A Unified Pixel-level Vision LLM for Understanding, Generating, Segmenting, Editing

awesome-vlm-architectures. Famous Vision Language Models and Their Architectures

aisuite. Simple, unified interface to multiple Generative AI providers

ASCII-generator. ASCII generator (image to text, image to image, video to video)

MiLM-6B.

Autoregressive-Models-in-Vision-Survey. [TMLR 2025🔥] A survey for the autoregressive models in vision.

DBO_code. This code implements the Dung Beetle Optimizer algorithm, inspired by the navigation behavior of dung beetles, to solve global optimization problems by achieving an adaptive balance between global and local search.

Awesome-CLIP. Awesome list for research on CLIP (Contrastive Language-Image Pre-Training).

SARDet_100K. [NeurIPS 2024 spotlight] Offical implementation of MSFA and release of SARDet_100K dataset for Large-Scale Synthetic Aperture Radar (SAR) Object Detection

LLaVA-NeXT-Reproduced. Reproduced LLaVA-NeXT with training code and scripts.

nanoInternLM. Jupyter Notebook

SA-UNet. The open source code of SA-UNet: Spatial Attention U-Net for Retinal Vessel Segmentation.

AI-Scientist. The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery 🧑‍🔬

ichigo. Local realtime voice AI

developer2gwy. 公务员从入门到上岸，最佳程序员公考实践教程

MiniCPM-V. A Pocket-Sized MLLM for Ultra-Efficient Image and Video Understanding on Your Phone

nanoGPT. The simplest, fastest repository for training/finetuning medium-sized GPTs.

CogVideo. text and image to video generation: CogVideoX (2024) and CogVideo (ICLR 2023)

paper-reading. 深度学习经典、新论文逐段精读

MindSearch. 🔍 An LLM-based Multi-agent Framework of Web Search Engine (like Perplexity.ai Pro and SearchGPT)

TensorRT-YOLO. 🚀 Easier & Faster YOLO Deployment Toolkit for NVIDIA 🛠️

echomimic. [AAAI 2025] EchoMimic: Lifelike Audio-Driven Portrait Animations through Editable Landmark Conditioning

FADC. CVPR 2024 Highlight: Frequency-Adaptive Dilated Convolution for Semantic Segmentation

MinerU. Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

PaddleMIX. Paddle Multimodal Integration and eXploration, supporting mainstream multi-modal tasks, including end-to-end large-scale multi-modal pretrain models and diffusion model toolbox. Equipped with high performance and flexibility.

xiaobei. 为 OPC/中小微企业老板们量身打造的"AI 搞钱搭子"

awesome-ai-coding. Awesome AI Coding

Multi-Scale-Attention. [JBHI] Code for our paper "Multi-scale Guided Attention for Medical Image Segmentation"

FunLLMTry. 一些对 LLM 有趣的测试

XuanYuan. 轩辕：度小满中文金融对话大模型

awesome-english-ebooks. 经济学人(含音频)、纽约客、卫报、连线、大西洋月刊等英语杂志免费下载,支持epub、mobi、pdf格式, 每周更新

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

Qwen3. Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.

GLM-4. GLM-4 series: Open Multilingual Multimodal Chat LMs | 开源多语言多模态对话模型

academicpages.github.io. Github Pages template based upon HTML and Markdown for personal, portfolio-based websites.

Prompt2Sign. [ICCV 2025 CV411y] The Data and Code of Prompt2Sign: A Comprehensive Multilingual Sign Language Dataset.

awesome-speech-recognition-speech-synthesis-papers. Automatic Speech Recognition (ASR), Speaker Verification, Speech Synthesis, Text-to-Speech (TTS), Language Modelling, Singing Voice Synthesis (SVS), Voice Conversion (VC)

LangGPT. LangGPT: Empowering everyone to become a prompt expert! 🚀 📌 结构化提示词（Structured Prompt）提出者 📌 元提示词（Meta-Prompt）发起者 📌 最流行的提示词落地范式 | Language of GPT The pioneering framework for structured & meta-prompt design 10,000+ ⭐ | Battle-tested by thousands of users worldwide Created by 云中江树

Agent-FLAN. [ACL2024 Findings] Agent-FLAN: Designing Data and Methods of Effective Agent Tuning for Large Language Models

InternLM-Math. State-of-the-art bilingual open-sourced Math reasoning LLMs.

InternEvo. InternEvo is an open-sourced lightweight training framework aims to support model pre-training without the need for extensive dependencies.

InternEvo_bak. InternEvo is a high-performance training system for giant models.

OpenAOE. LLM Group Chat Framework: chat with multiple LLMs at the same time. 大模型群聊框架：同时与多个大语言模型聊天。

pytorch_sphinx_theme. PyTorch Sphinx Theme

.github.