An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale
8 files for Claude Code, OpenCode and Codex: algorithm-expert, fsdp-expert, launcher-expert, megatron-expert and 4 more — 2,533 tokens loaded in every session.
AGENTS.md A 1,627 tok CLAUDE.md A 906 tok .opencode/commands/commit.md A 7 tok .opencode/agents/algorithm-expert.md A 37 tok .opencode/agents/fsdp-expert.md A 38 tok .opencode/agents/launcher-expert.md A 38 tok .opencode/agents/megatron-expert.md A 29 tok .opencode/agents/ray-expert.md A 56 tok These files are redai-infra/Relax's own configuration — they tell Claude Code, OpenCode and Codex how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.