Miles

license

Apache-2.0(OSI)

source

public(radixark/miles)

core features withheld

no — no enterprise or ee path in the repository root and no paid tier described in the README

Apache-2.0, an OSI license, with the full source public in this repository. There is no enterprise or ee path in the repository and no paid tier described in the README.

• https://raw.githubusercontent.com/radixark/miles/main/LICENSE recorded 2026-10-07

  LICENSE body, read in full: the stock Apache License 2.0 text through "END OF TERMS AND CONDITIONS" and the unfilled template appendix, with no appended term or additional condition.

• https://raw.githubusercontent.com/radixark/miles/main/README.md recorded 2026-10-07

  README, read for the distribution and gating question: "Miles is a high-performance, enterprise-ready reinforcement learning framework for" large-scale post-training. It points to installing and running the code from this repository ("Install Miles", "Quick Start"); its "enterprise-ready" wording describes the framework, and it names no enterprise edition, pricing, paid or licence-gated build.

No package registry carries this product, so GitHub stars are the only available adoption signal. A star is not a use, so the count keeps confidence low; a PyPI project also named `miles` exists but is an unrelated, undocumented package with no link to this repository, so it is not counted here.

• https://api.github.com/repos/radixark/miles recorded 2026-08-31

  Repo metadata for radixark/miles: stargazers_count 2293, archived false, pushed_at 2026-08-31.

Miles's asynchronous design and P2P RDMA weight transfer put it in the same infrastructure class as verl, at the same trillion-parameter scale. It sits a step below verl because verl publishes a substantially wider algorithm catalog, while Miles documents only PPO and GRPO.

• https://raw.githubusercontent.com/radixark/miles/main/README.md recorded 2026-08-31

  About and Performance sections describe pairing SGLang for rollout with Megatron-LM for scalable training, a PyTorch FSDP2 backend alternative, fully async RL with decoupled rollout/training workers, and in-loop weight updates over P2P RDMA on a trillion-parameter model (Kimi-K2.6).