菜单
How `max_env_worker` caps the "Running Episodes" gauge, and how `AgentJetJob` relates to the YAML config.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/modelscope/AgentJet --skill swarm-configuration命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Execute AgentJet reinforcement learning experiments using experiment blueprints in classic (non-swarm) mode. Handles full lifecycle: launch experiment in tmux, monitor progress, analyze errors, collect results, and write finish flag. Use when the user wants to run AgentJet training experiments without the swarm distributed framework.
Execute AgentJet reinforcement learning experiments using experiment blueprints in swarm mode. Handles full lifecycle: generate blueprint if needed, launch experiment in tmux, monitor progress, analyze errors, collect results, and write finish flag. Use when the user wants to run or debug AgentJet training experiments.
Install AgentJet swarm server using Conda. Handles Python 3.10 environment creation, dependency installation with the verl training backbone, flash-attn compilation, and optional PyPI mirror for China users.
基于 SOC 职业分类
正在显示 SKILL.md
| name | swarm-configuration |
| description | How `max_env_worker` caps the "Running Episodes" gauge, and how `AgentJetJob` relates to the YAML config. |
| license | Complete terms in LICENSE.txt |
The Running Episodes (Episodes: N) number in swarm_overwatch is bounded by the engine-side max_env_worker (set on the job config, e.g. CocktailV2Config.max_env_worker, then forwarded into AgentJetJob and read at ajet/backbone/trainer_verl.py as max_parallel). In ajet/task_rollout/native_parallel_worker.py::rollout_swarm, the engine spawns ceil(max_env_worker / grpo_n) * grpo_n long-lived worker threads, each looping register_episode → wait-for-claim → repeat, so the total in-flight episodes (summed across all swarm clients) cannot exceed that count. total_batch_size, per-client max_env_worker, grpo_n, and the number of clients do not raise this cap , to lift it, raise the engine's max_env_worker (keep it divisible by grpo_n) and restart.
When using Agentjet Swarm, please first use AgentJetJob as the primary configuration interface.
If there are fields you want to set that are not exposed as AgentJetJob kwargs, use yaml as the primary configuration interface.
In general, you should place most configuration in a place (either AgentJetJob or yaml), and MUST NOT place configuration here and there at the same time.
AgentJetJob (ajet/copilot/job.py) is a thin YAML overlay, not a separate config system. On __init__ it loads a base YAML (default ajet/default_config/ajet_swarm_default.yaml, or whatever path is passed via base_yaml_config=) into self.config, then walks an overrides table that maps each constructor kwarg to a deep YAML key (e.g. max_env_worker → ajet.rollout.max_env_worker, batch_size → ajet.data.train_batch_size, model → ajet.model.path). For each entry: if the kwarg is None the YAML value wins; if non-None it overwrites the YAML value in-place. Anything not listed in overrides (e.g. rollout.temperature, rollout.multi_turn, trainer_common.save_freq) has no kwarg shortcut and must be set by mutating ajet_job.config.ajet.* directly after construction , this is what build_cocktail_ajet_job does in the cocktail_rl_v2 tutorial. dump_job_as_yaml(path) serialises the merged result back out, and that dumped YAML is the file the engine subprocess actually consumes. Net effect: YAML is the source of truth for defaults; AgentJetJob kwargs are sparse overrides; post-construction attribute writes are the escape hatch for fields without a kwarg.
