- China
-
15:15
(UTC +08:00)
Starred repositories
Portable C++17 implementation of ACE-Step 1.5 music generation using GGML. Text + lyrics in, stereo 48kHz MP3 or WAV out. Runs on CPU, CUDA, ROCm, Metal, Vulkan.
A unified library of SOTA model optimization techniques like quantization, pruning, distillation, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks …
Pure C++ implementation of several models for real-time chatting on your computer (CPU & GPU)
finetune method to create think/model/requires tags to allow LLMs to write programs for things they can calculate instead of hallucinating answers.
All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.
Videodl: A lightweight video downloader written in pure python. (轻量级视频下载器,优先高清无水印,支持抖音,快手,小红书,B站,TikTok,YouTube,FIFA+,优酷,腾讯,爱奇艺,1905电影网,乐视,芒果,咪咕,PPTV,搜狐,Facebook,Twitter,新浪微博,今日头条,网易公开课,全民K歌,CCTV央视…
A multi core friendly rigid body physics and collision detection library. Written in C++. Suitable for games and VR applications. Used by Horizon Forbidden West and Death Stranding 2.
High-performance ACID relational database in C++ with MVCC, B+ tree indexes, and SQL query optimization
[CVPR 2026] Fast-FoundationStereo: Real-Time Zero-Shot Stereo Matching
Is it difficult to develop C++ high-concurrency server applications? Come and use XServer
On-device AI across mobile, embedded and edge for PyTorch
LiteRT, successor to TensorFlow Lite. is Google's On-device framework for high-performance ML & GenAI deployment on edge platforms, via efficient conversion, runtime, and optimization
A lightweight, fast JavaScript runtime built on QuickJS and libuv
Rust multi‑backend OCR/VLM engine (DeepSeek‑OCR-1/2, PaddleOCR‑VL, DotsOCR) with DSQ quantization and an OpenAI‑compatible server & CLI – run locally without Python.
New generation of CLIP with strong fine grained discrimination capability, ICML2025
Simple MPL-2.0-licensed C++ geometry processing library.
Transforms complex documents like PDFs into LLM-ready markdown/JSON for your Agentic workflows.
[CVPR 2025] DEIM: DETR with Improved Matching for Fast Convergence
A library for building fast, reliable and evolvable network services.
DEPRECATED. Evmongoose is an asynchronous, event(libev) based multi-protocol embedded networking library with functions including TCP, HTTP, WebSocket, MQTT and much more. It's based on mongoose an…
CommonMark/Markdown Java parser with source level AST. CommonMark 0.28, emulation of: pegdown, kramdown, markdown.pl, MultiMarkdown. With HTML to MD, MD to PDF, MD to DOCX conversion modules.
A high-performance runtime framework for modern robotics.
LightSeq: A High Performance Library for Sequence Processing and Generation
[ECCV'24] Kalman-Inspired Feature Propagation for Video Face Super-Resolution