An open-source solution for full parameter fine-tuning of DeepSeek-V3/R1 671B, including complete code and scripts from training to inference, as well as some practical experiences and conclusions. (DeepSeek-V3/R1 满血版 671B 全参数微调的开源解决方案，包含从训练到推理的完整代码和脚本，以及实践中积累一些经验和结论。)

Ollama负载均衡服务器 | 一款高性能、易配置的开源负载均衡服务器，优化Ollama负载。它能够帮助您提高应用程序的可用性和响应速度，同时确保系统资源的有效利用。

LLMVoX: Autoregressive Streaming Text-to-Speech Model for Any LLM

这是一个高性能的LLM推理API，且自带UI界面，它将DeepSeek R1 的思维推理信息和Anthropic的 Claude 系列模型相集成。

Complex Multimodal Reasoning with Reinforced LLMs/VLMs/MLLMs. Motivated by the DeepSeek-R1-Zero

Convert LLaMA3.1-8B to DeepSeek R1 MLA & MoE (raw)

Llama post-training examples with torchtune

Speech Resynthesis and Language Modeling with Flow Matching and Llama

Unified API platform for free access to enterprise-grade AI models from Google, Groq, and OpenRouter. Industrial-ready integration with high-performance Models Inc. DeepSeek-R1, QwQ 32B

Chat with a Codebase using DeepSeek R1, Claude, GPT, Gemini

The official code of paper Critic-V: VLM Critics Help Catch VLM Errors in Multimodal Reasoning

Reproduction of the complete process of DeepSeek-R1 on small-scale models, including Pre-training, SFT, and RL.

AgentMake AI: a kit for developing agentic AI applications that support 14 AI backends and and work with 7 agentic components, such as tools and agents. (Developer: Eliran Wong) Supported backends: anthropic, azure, cohere, custom, deepseek, genai, github, googleai, groq, llamacpp, mistral, ollama, openai, vertexai, xai

See How Big Exactly A 128k Token Text Is

全栈式AI服务平台技术方案，支持 DeepSeek，OpenAI，Midjourney，Stable Diffusion，DALL·E，Gemini等模型，支持多端使用，自定义模型对接，对话会话隔离、云端存储、支持多设备共享数据，多模态模型支持，支持图片解析，支持联网搜索，支持AI智能体应用，支持官方微信支持、易支付等支付方式，数据统计, 丰富美观的仪表盘统计。

DeepSeek 中文官网、网页版入口、下载及使用教程 | DeepSeek 最全教程指南~ （2025年3月更新） 轻松使用 DeepSeek 网页版，享受 快速、稳定、不卡顿 的体验，支持 DeepSeek R1 满血版~ 本指南提供全面的 DeepSeek 使用教程指南，包括 DeepSeek 官网入口、DeepSeek 网页版、DeepSeek 下载 和 DeepSeek 平替网站，助力顺畅使用 DeepSeek~

Integrates OVMS V3 into Home Assistant through HACS. Communicates via MQTT, automatically creating and parsing vehicle entities. Developed with AI assistance.

AM1 - Your own web application to do DeepResearch using any Multi-Agent Framework and underlying LLM to run on production servers or your laptop.

Serverless DeepSeek R1 Inference with FastAPI and Lambda SnapStart

GLM深远海船舶作业大模型应用赛——个人技术方案 "多Agent协同系统“

Stable Diffusion WebUI

Данный проект демонстрирует базовую реализацию смены лица с использованием Stable Diffusion Inpainting Pipeline. Основная идея – заменить лицо на целевом изображении так, чтобы оно выглядело фотореалистично и органично, с сохранением узнаваемых черт, аналогичных референсному лицу.

Ollama中llama.cpp的Go语言绑定

well this is one of my good project

LexiSynapse es un asistente de estudio basado en inteligencia artificial, diseñado para optimizar la organización académica y mejorar la eficiencia en el aprendizaje. Desarrollado en Python con principios de Programación Orientada a Objetos (POO)

Minimal script I'm using to train DeepSeek R1 1.5B on a RTX4060 8GB

This script demonstrates how to fine-tune a distilled Llama model using Unsloth for medical question answering.

Chatty AI transforms your terminal into a vibrant hub where history's greatest minds come alive. Chat one-on-one with Einstein about relativity, watch Shakespeare and Isaac Newton debate the human condition, or let Marie Curie and Ada Lovelace discuss women in science.

llama-server in Docker.

LLM-based Q&A on preloaded docs, raw data, Wikipedia articles and scraped web pages with knowledge graphs, analytics, charts and Streamlit interface

欢迎体验基于**RAG（检索增强生成）**架构的智能问答系统！📚 这个小助手不仅能读懂你的PDF，还能像个聪明的小伙伴一样回答你的问题！✨ 我们用Ollama作大脑，通过向量检索让文档问答变得更聪明、更高效！🧠

A web-based application for generating AI-powered images using Stable Diffusion.

Use RAG to build web app to recommend trending movies based on users' moods. Tech stacks: TMBD API, Python, LangChain, GPT-4o-mini

Generate Pokémon images from text with Stable Diffusion and an interactive Gradio interface.

PickHacks - Hackathon2025 || Winner - 3rd Place Overall & Winner - Best use of Gemini API

A practical Python reference manual covering everything from data wrangling to machine learning, deep learning, and AI—complete with ready-to-use code snippets, interactive examples, and downloadable Jupyter notebooks.

一个利用LLM能力进行翻译并自动修订翻译结果的工具，可视化展示修订前后的差异。A tool that leverages LLM capabilities to translate and automatically revise translation results, visualizing the differences before and after revision.

run llama.cpp quickly and conveniently.

A Very powerful summarization tool that processes and extracts key insights from content using Llama 2, Haystack, Whisper, and Streamlit. Runs smoothly on CPU with Llama

Deepseek R1 and Qdrant-Powered Financial Doc Intelligence

本地部署DeepSeek-R1接入微信机器人，实现个性化微信助手

This project integrates Segment Anything Model (SAM) with Stable Diffusion and LaMa for mask-free image inpainting. Users can click on objects to remove, fill, or replace them with AI-generated content.

Link download👇

Maintenance of a set of tools to enhance LLM through MCP protocols.

Local RAG implementation using Deepseek R1

Skin Condition Analysis using YOLOv8, InceptionV3, and Qwen2-VLM

The Next Step Forward in Multimodal LLM Alignment

A comprehensive guide for running Large Language Models on your local hardware using popular frameworks like llama.cpp, Ollama, HuggingFace Transformers, vLLM, and LM Studio. Includes optimization techniques, performance comparisons, and step-by-step setup instructions for privacy-focused, cost-effective AI without cloud dependencies.

VSCode extension that lets you run LLM DeepSeek-R1 offline/locally, based on a Beyond Fireship video.

Consolidates and de-duplicates facts automatically generated by Bee AI using k-means clustering and a locally hosted AI hosting using LM Studio.

Implementation of Group Relative Policy Optimization (GRPO) to fine-tune Open Language Models like LlaMa-3.2, Qwen2 for Math Tasks.

LlamaEngine – A cross-platform C++ library for dynamically loading and running llama.cpp with CPU, Vulkan, Metal, and CUDA backends.

This uses Deepseek-R1 and Ollama to improve your writing.

Agentic AI Stock Analysis is an AI-powered multi-agent system that analyzes stock market trends, provides analyst recommendations, and fetches the latest news using Groq AI models. It integrates financial and web search agents to deliver real-time insights.

基于DeepSeek-R1、bge-m3、bge-reranker-v2-m3、dify技术框架实现中文本地知识库，依托项目搭建，搭建中文大模型知识库LLM支持多模态、侧端

Aici aveti materialele pentru tutoriatul la LMC (2024-2025)

A specialized question-answering system focused on medieval arms, armor, and related topics using a fine-tuned LLaMA 3.2 model.

Ada is a cyberpunk AI girlfriend powered by LLaMA 3 70B. Ada provides a real-time, ultra-responsive, immersive conversations in a high-tech, dystopian-inspired interface. Ada adapts effortlessly for engaging discussions or immersive roleplay. Chat with Ada today✨!

Simple configurable wrapper for local LLM-based test refinery over llama.cpp API

Volatility Forecasting of CSI 1000 Index using GARCH(1,1) coding by DeepSeek R1

完全使用HTML实现的AI对话，兼容大部分api

Unreal Engine Projects Shortcut Manager

🚀MasterMind – A smart learning platform designed to **enhance your study experience** with curated courses, interactive materials, and an intuitive dashboard. Stay organized, learn efficiently, and master new skills effortlessly. 💡📚

Llama2 Fine-tuned on MindsDB Docs

A library of simple HTML website's generated by DeepSeek-R1

Tool to give AI generated score, to analyze with patterns how much input text is AI generated using AWS Bedrock Llama 3.1 405B

DeepSeek Think+Gemini = DeepGemini

Colaboratory上でQwen2.5-VLをお試しするサンプル

ML estimation of scale mixture of skew-normal linear mixed models Use smsn.lmm (skewlmm) With (In) R Software

AI-powered image analysis and slide deck generator with OpenAI, Gemini, and Claude support

AI-Powered Blog Platform is a modern blogging application built with Next.js, Tailwind CSS, Firebase, and the Gemini API. It enables AI-assisted content creation, real-time database management, and secure authentication, all within a fully containerized Docker environment for seamless deployment and scalability.

Extract structured data from PDF documents using Google Gemini AI. This tool converts PDFs into clean JSON without predefined schemas, ideal for invoices, forms, and tables.