Stars
DeepSeek Harness: Everything is a Plugin.
Generate production-quality SVG+PNG technical diagrams from natural language. 7 styles, UML support, and AI/Agent workflow patterns.
A high-throughput and memory-efficient inference and serving engine for LLMs
Practical quantization recipes for large language models and speech models, from model preparation through deployment-oriented validation.
Real-time text-to-speech with Qwen3-TTS
LLM model quantization (compression) toolkit with HW acceleration support for Nvidia, AMD, Intel GPU and Intel/AMD/Apple CPU via HF, vLLM, and SGLang.
Native End-to-End Full-Duplex Spoken Language Model
Codex skill for converting slide images, PDFs, and image-based PPTX files into editable PowerPoint decks.
AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI
Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
GPT-2 training in pure Mojo with hand-written CUDA and Metal GPU kernels. llm.c parity in bf16 on NVIDIA, 1.72x faster than PyTorch MPS on Apple Silicon.
Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
A framework for efficient model inference with omni-modality models
10000 chatTTS voices ！chatTTS 音色库，再也不为音色抽卡烦恼啦。这是我第一个项目，熬夜龟速生产10000条音色并上传Github，给点鼓励呗哈！主域名：https://www.TTSlist.com 备用：http://ttslist.aiqbh.com/
Qwen3-ASR is an open-source series of ASR models developed by the Qwen team at Alibaba Cloud, supporting stable multilingual speech/music/song recognition, language detection and timestamp prediction.
解决 Cursor 使用 DeepSeek V4 模型时的 `reasoning_content must be passed back` 错误
Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞
A theoretical reconstruction of the Claude Mythos architecture, built from first principles using the available research literature.
Verbose implementations of LLMs architectures, techniques and research papers from scratch. Qwen..., RLHF, Multimodal, Hyperconnections...
FastAPI-compatible Python framework with Zig HTTP core; 7x faster, free-threading native
Kyutai's Speech-To-Text and Text-To-Speech models based on the Delayed Streams Modeling framework.
AirLLM 70B inference with single 4GB GPU
Train transformer language models with reinforcement learning.
A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.