A service-based, data-driven asynchronous RL engine for LLMs — no central controller, no Ray.
Meshy models every role of an RL run as an independent service. Samples flow between services through a single TransferQueue data plane, control flow is driven by data availability, and the whole topology is derived locally by each process from one declarative recipe. Built on SGLang and torchtitan.
Content type
Image
Digest
sha256:4b99fa423…
Size
13.6 GB
Last updated
21 days ago
docker pull ztonyzhao/meshy:0.1.0-alpha