Stars
SGLang is a fast serving framework for large language models and vision language models.
A high-throughput and memory-efficient inference and serving engine for LLMs
A generative world for general-purpose robotics & embodied AI learning.
Ongoing research training transformer models at scale
Train transformer language models with reinforcement learning.
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
A PyTorch native platform for training generative AI models
Build resilient language agents as graphs.
Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek-R1, Qwen3, Gemma 3, TTS 2x faster with 70% less VRAM.
A Flexible Framework for Experiencing Cutting-edge LLM Inference Optimizations
DeepEP: an efficient expert-parallel communication library
Trainable fast and memory-efficient sparse attention
🚀 Efficient implementations of state-of-the-art linear attention models
Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.
✨✨Latest Advances on Multimodal Large Language Models
Leaderboard Comparing LLM Performance at Producing Hallucinations when Summarizing Short Documents
Distribute and run LLMs with a single file.
[TMLR] A curated list of language modeling researches for code (and other software engineering activities), plus related datasets.
Set of tools to assess and improve LLM security.
RAGEN leverages reinforcement learning to train LLM reasoning agents in interactive, stochastic environments.
Run your own AI cluster at home with everyday devices 📱💻 🖥️⌚
gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI
An Easy-to-use, Scalable and High-performance RLHF Framework based on Ray (PPO & GRPO & REINFORCE++ & vLLM & Ray & Dynamic Sampling & Async Agentic RL)
A throughput-oriented high-performance serving framework for LLMs
[CVPR 2023 Best Paper Award] Planning-oriented Autonomous Driving
General Reasoner: Advancing LLM Reasoning Across All Domains [NeurIPS25]