Stars
Reference PyTorch implementation and models for DINOv3
A reproduction of the Deepseek-OCR model including training
Official implementation of "Continuous Autoregressive Language Models"
MNN is a blazing fast, lightweight deep learning framework, battle-tested by business-critical use cases in Alibaba. Full multimodal LLM Android App:[MNN-LLM-Android](./apps/Android/MnnLlmChat/READ…
A lightweight suite of motion imitation methods for training controllers.
Batch processor to enable large content be digested by Ollama, focused around book processing and translations by default, fully, configurable through json.
High-performance, modular vehicle physics system for Unreal Engine 5, fully implemented in C++ with real-time tunable parameters. (made in UE5.3) Now supports multiplayer and level squence(if axles…
Welcome to the official repository of SINQ! A novel, fast and high-quality quantization method designed to make any Large Language Model smaller while preserving accuracy.
⭐All-in-one AI Companion! AI Desktop Companion + AI Virtual Streamer + AI Social App Bot + AI Interactive UI Interface + Digital Human Broadcasting + AI Games and all the features you can imagine! …
Kilo is the all-in-one agentic engineering platform. Build, ship, and iterate faster with the most popular open source coding agent. #1 on OpenRouter. 750k+ Kilo Coders. 6.1 trillion tokens/month.
⏩ Ship faster with Continuous AI. Open-source CLI that can be used in TUI mode as a coding agent or Headless mode to run background agents
Wan: Open and Advanced Large-Scale Video Generative Models
An app that brings language models directly to your phone.
对比测试不同大语言模型(LLM)性能的工具平台，支持DeepSeek API、Ollama本地模型和VLLM本地模型。A simple tools to test multi models and display the time cost.
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+。
天若ocr开源版本的本地版，采用Chinese-lite和paddleocr识别框架
Hierarchical Reasoning Model Official Release