My own open source implementation of OpenAI's new Deep Research agent. Get the same capability without paying $200. You can even tweak the behavior of the agent with adjustable breadth and depth. Run it for 5 min or 5 hours, it'll auto adjust.

Witness the aha moment of VLM with less than $3.

Keep searching, reading webpages, reasoning until it finds the answer (or exceeding the token budget)

Solve Visual Understanding with Reinforced VLMs

This repo contains the dataset and code for the paper "SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering?"

An AI chat bot for small and medium-sized teams, supporting models such as Deepseek, Open AI, Claude, and Gemini. 专为中小团队设计的 AI 聊天应用，支持 Deepseek、Open AI、Claude、Gemini 等模型。

推荐使用DeepSeekV3。可以接入微信、QQBot。基于LLM的更逼真的情感陪伴程序。内置了 Atri-My dear moments 的 prompt。Built-in prompt for Atri My dear moments.Support WeChat& QQBot. More realistic emotional companionship program based LLM, meet the characters in your dream.

(Supports DeepSeek R1) An AI-powered research assistant that performs iterative, deep research on any topic by combining search engines, web scraping, and large language models.

Reliable Automation Agents at Scale

AI Pair Programming in Emacs

基于DeepseekR1/Gemini2/GoogleVison2/GoogleSearch API实现的深度思考联网多模态大模型。

Extend OpenRLHF to support LMM RL training for reproduction of DeepSeek-R1 on multimodal tasks.

Local LLM Powered Recursive Search & Smart Knowledge Explorer

Dive is an open-source AI Agent desktop application that seamlessly integrates any Tools Call-supported LLM with frontend MCP Server—part of the Open Agent Platform initiative.

Community maintained hardware plugin for vLLM on Ascend

[arXiv] On-device Sora: Enabling Diffusion-Based Text-to-Video Generation for Mobile Devices

🔬 Visualize attention layers from Stable Diffusion

Run Stable Diffusion on Android Devices with Snapdragon NPU acceleration. Also supports CPU inference.

SOTA RL fine-tuning solution for advanced math reasoning of LLM

fine-tune deepseek r1

Flame is an open-source multimodal AI system designed to translate UI design mockups into high-quality React code. It leverages vision-language modeling, automated data synthesis, and structured training workflows to bridge the gap between design and front-end development.

AIN - The First Arabic Inclusive Large Multimodal Model. It is a versatile bilingual LMM excelling in visual and contextual understanding across diverse domains.

Demonstrates how to build a streaming enterprise agent using Azure AI Agent Service. This sample integrates local HR and company policy documents, Bing for external context, and GPT-4o (though the service is model-agnostic) to deliver real-time Q&A and automated workflows.

Super-easy AI Agent Library specialized for LLM function calling

Using the moondream VLM with optical flow for promptable object tracking

Enhanced BurpGPT 是一个强大的 Burp Suite 插件。通过分析指定的 HTTP 请求和响应，帮助安全测试人员更快速地发现潜在的安全漏洞。

A novel alignment framework that leverages image retrieval to mitigate hallucinations in Vision Language Models.

An AI-powered tool for summarizing YouTube videos by generating scene descriptions, translating them, and creating subtitled videos with text-to-speech narration

GRPO for VLMs

Enable your ollama model to participate in online chats.

[ICLR 2025] - Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion

Securade.ai Sentinel - A monitoring and surveillance application that enables visual Q&A and video captioning for existing CCTV cameras.

ComfyUI-DeepSeek-Toolkit is a node developed within the ComfyUI framework, designed to bring powerful LLMs like DeepSeek and Qwen into practical ComfyUI development. The goal is to leverage the capabilities of these advanced LLMs to create a series of useful nodes for ComfyUI.

The official code of "PixelWorld: Towards Perceiving Everything as Pixels"

A benchmarking tool for comparing different LLM API providers' DeepSeek model deployments.

micro llm pt/sft/grpo from scratch

an open-source, community-curated directory of contexts for use with LLMs

Meet WebHive, the AI-powered browser that takes care of tasks for you. No more endless clicks, tell it what you need, and it gets it done.

AI ChatBox是一个基于MFC框架开发的本地AI聊天程序，集成了大型语言模型，支持多轮对话，对话记录保存等功能。用户可以选择不同的本地模型，例如qwen、llama、deepseek等，进行对话。AI ChatBox完全离线运行，保护用户隐私

Chitrarth: Bridging Vision and Language for a Billion People

用于astrbot，使用Stable Diffusion WebUI的API进行图片生成的插件。

I'll be your machinery.