Miles system makes advanced reinforcement learning easier and more reliable
Miles v0.1: Production-Level Post-Training
Machine LearningComputation and Language
Summary
Training advanced AI systems often requires complex setups that can be hard to manage and scale. The creators of Miles introduced a fully functional system designed to simplify and improve the training process for reinforcement learning, focusing on accuracy, efficiency, and flexibility. The system includes tools for running training steps, synchronizing model updates, and supports various training methods. They showed how Miles can train a large AI model effectively using many graphics processors, demonstrating its reliability and speed. The system is open-source, allowing others to use and build upon it.
reinforcement learningpost-trainingrollout enginetrainer backendweight synchronizationLoRAon-policy distillationdiffusion modelsNVIDIA Megatron-LMPyTorch FSDP
Authors
RadixArk, :, Tom Chen, Mao Cheng, Shi Dong, Kangrui Du, Yanbin Jiang, Jiajun Li, Yiming Li, Tao Lin, Yusheng Su, Andy Ye, Yueming Yuan, Zhichen Zeng
Abstract
We present Miles v0.1, a full-stack, production-ready system for frontier post-training. Building upon the clean design of slime, Miles designs each stage of the reinforcement-learning (RL) training loop around a single principle: components should be verified, clean, and customizable. With accuracy, efficiency, reliability, and scalability as first-class goals, Miles aims to make frontier-scale RL accessible to researchers and enterprises alike. This report walks through the system end to end: rollout engines built on SGLang, a trainer with a choice of two backends (NVIDIA Megatron-LM and PyTorch FSDP), and three weight-synchronization transports for different deployment topologies. Beyond full-parameter RL, Miles also supports LoRA RL, on-policy distillation, supervised fine-tuning, and true-on-policy rollout-training alignment, and extends the same architecture to diffusion models. We close with an end-to-end case study: fully asynchronous agentic RL on a GLM-5.2 744B-A40B model over terminal-use coding tasks, running on 64 NVIDIA GB300 GPUs with a median step time of 263 seconds over the first 30 measured steps. Miles is open-sourced at https://github.com/radixark/miles, with the project website at https://miles.radixark.com.