Living memory for AI

Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.

Open-source implementation of AlphaEvolve

Train your Agent model via our easy and efficient framework

A GPU cluster manager that configures and orchestrates inference engines like vLLM and SGLang for high-performance AI model deployment.

Production-ready platform for agentic workflow development.

A Python library for extracting structured information from unstructured text using LLMs with precise source grounding and interactive visualization.

Fast Multimodal Semantic Deduplication & Filtering

Puffing up reinforcement learning

Agentic RL Training at Scale

Our library for RL environments + evals

SkyRL: A Modular Full-stack RL Library for LLMs

Port of OpenAI's Whisper model in C/C++

🤗 LeRobot: Making AI for Robotics more accessible with end-to-end learning

DFloat11 [NeurIPS '25]: Lossless Compression of LLMs and DiTs for Efficient GPU Inference

RAGEN leverages reinforcement learning to train LLM reasoning agents in interactive, stochastic environments.

Machine Learning Engineering Open Book

📰 Must-read papers on KV Cache Compression (constantly updating 🤗).

SOTA search powered LLM

Official Implementation of "KBLaM: Knowledge Base augmented Language Model"

Stateful LLM Serving

This is the homepage of a new book entitled "Mathematical Foundations of Reinforcement Learning."

GPU documentation for humans

Network Analysis in Python

[NeurIPS'25] Official codebase for "SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution"

Muon is an optimizer for hidden layers in neural networks

FlashMLA: Efficient Multi-head Latent Attention Kernels

Materials for learning SGLang

Fully open reproduction of DeepSeek-R1

[NeurIPS 2025 Spotlight] Reasoning Environments for Reinforcement Learning with Verifiable Rewards