cuda-oxide is an experimental Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign languag…

OxiMedia is the Sovereign Media Framework - A patent-free, memory-safe multimedia processing library written in pure Rust. Pure Rust reconstruction of both FFmpeg (multimedia processing) and OpenCV…

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, Andr…

Free and open source video editor, based on MLT Framework and KDE Frameworks

Tiny, no-nonsense, self-contained, Tensorflow and ONNX inference

Inference at the speed of light.

Qwen3.5-Thor — High-performance BF16/NVFP4 inference engine for Qwen3.5 model family on NVIDIA Jetson AGX Thor (SM110a Blackwell). C++17/CUDA, Ollama/OpenAI compatible API.

Uaha is built on the Aha engine, a desktop application constructed with native technologies, designed to manage AI models and API services through a modern user interface.

音色克隆工具

API + AI 融合网关

A CLI tool for downloading models from Modelscope

Service configuration and registry center implemented in Rust, Like Nacos

MNBVC(Massive Never-ending BT Vast Chinese corpus)超大规模中文语料集。对标chatGPT训练的40T数据。MNBVC数据集不但包括主流文化，也包括各个小众文化甚至火星文的数据。MNBVC数据集包括新闻、作文、小说、书籍、杂志、论文、台词、帖子、wiki、古诗、歌词、商品介绍、笑话、糗事、聊天记录等一切形式的纯文本中文数据。

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

[NeurIPS 2024] Depth Anything V2. A More Capable Foundation Model for Monocular Depth Estimation

[CVPR 2025 Best Paper Nomination] FoundationStereo: Zero-Shot Stereo Matching

使用深度学习对人体心电数据进行多分类

CycleResearcher: Improving Automated Research via Automated Review

[CVPR 2025] RoboBrain: A Unified Brain Model for Robotic Manipulation from Abstract to Concrete. Official Repository.

RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs

从小说中提取对话数据集

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

[EMNLP 2024] 中文领域心理健康对话大模型MeChat

Train transformer language models with reinforcement learning.

Course to get into Large Language Models (LLMs) with roadmaps and Colab notebooks.

本项目旨在分享大模型相关技术原理以及实战经验（大模型工程化、大模型应用落地）

利用HuggingFace的官方下载工具从镜像网站进行高速下载。

Mamba SSM architecture

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