← Discover
Discover · GitHub · Miles↑ +64 today

Enterprise-Scale AI Training

Miles is an enterprise-focused framework that manages reinforcement learning processes for large language models (LLMs) and vision-language models (VLMs). Developed to optimize post-training processes, this Python-based tool evolves in parallel with the slime project.

Updates

  • September 5, 2026: Stars 2,601 → 2,603, latest release v0.1.0 (August 18, 2026).

What you get

  • Provides high-performance reinforcement learning processes for trillion-parameter models.
  • Offers integration with SGLang (high-speed model serving software) and Megatron-LM (scalable training library).
  • Features fault tolerance for automatic recovery without stopping training during system crashes.

Getting started

Related dictionary terms

Who it is forSuitable for enterprise developers and researchers looking to train and optimize large-scale AI models.
LicenseApache-2.0

Links

TreScout did not build this tool · we found it in GitHub trends and wrote it up. This page describes the repository as of 2026-09-05: The star count and our text belong to that day, the repository may have changed since. Check the repository link for the current state. This page was machine-translated from the Turkish original · the Turkish version prevails.